II een fi YES 0 NO El 3 I 0 IC CI AUTHOR SOURCE CODE ...Mar 69 II een fi 3 rbil 1 E. N.) fi \ii...

$: II een fi YES 0 NO El 3 I 0 IC CI AUTHOR SOURCE CODE ...Mar 69 II een fi 3 rbil 1 E. N.) fi \ii AUTHOR McDonald, Frederick J.; Koran, Mary Lou TITLE. The Effect of Individual. Differences$
OE FORM 6000, 2/69 DEPARTMENT OF HEALTH, EDUCATION, AND WELFAREOFFICE OF EDUCATION

ERIC ACC. NO.ED 032 291

mrtn,.. rumr-vrc I nr.=IJIVIC.

IS DOCUMENT COPYRIGHTED? YES 0 NO 1:1.. FO REPRODUCTION RELEASE? YES 0 NO El(IL 61 EL OF AVAILABILITYL I TY I 0 IC "CI

CH ACC. NO.

SP 003 095

P.A.

24

PUBL. DATE

Mar 69

II een fi 3

rbil 1 E. N.) fi \ii

AUTHOR

McDonald, Frederick J.; Koran, Mary Lou.TITLE

The Effect of Individual. Differences on Observational Learning in theAcquisition of a Teaching Skill. Final Report.

SOURCE CODE INSTITUTION ( BOURG E)CI02500 Stanford Univ., Calif., School of Education

SP. AG. CODE SPONSORING AGENCY

MO 004 Office of Education (DREW), Washington, D.C., Bureau of Research'CDR S PRICE CONTRACT NO. G$ANT NO.1.00;11.15

OEG- -8-0810 -011 -010REPORT NO.BUR EAU NO.

BR-8-1-073AVAILABILITY

JOURNAL CITATION

DESCRIPTIVE NOTE

221p..

.

DESCRIPTORS.

.

*Academic Aptitude; *Teaching Skills;.*Training Techniques; Education Majors;Films; Individual Differences; Learning; Microteaching; Teacher Characteristics;Teaching Models

IDENTIFIERS

. .

ABSTRACTA study examined the effects of verbal and perceptual dimensions of individualdifferences in relation to the efficacy of two different kinds of modelingprocedures in the acquisition of a teaching skill (analytic questioning).Aptitude tests for cognitive factors plus specially developed audiovisual testswere administered to 121 intern teachers randomly assigned to three treatmentgroups: a film-mediated modeling treatment (a filmed portrayal of analyticquestioning); a written modeling treatment (a text ofthe film sound track); and acontrol treatment which received no model, but went through all other stepsilcluding initial instructions and microteaching pretest and two cycles of models,rehearsal, and microteaching. The criterion performances assessed by trainedraters included the frequency, variety, and quality of analytic questions used inthree separate teaching sessions in addition to scores on two written posttests.Instructional treatment main effects as well as aptitude by treatment interactionswere investigated using analysis of variance and comparison of regression slopes.Findings, which supported hypotheses, suggest that the rate and level of learningof a specific teaching strategy varies as a function of model presentationfilm-mediated modeling most effective: no modeling least effective); and that the,effectiveness of instructional methods varies from S to S with such differencesbeing related to trainee aptitudes. (ED 017 985 and ED 028 982 are relateddocuments.) (JS)

0I0 O e7oa

U.S. DEPARTMENT OF HEALTH, EDUCATION & WELFARE

OFFICE OF EDUCATION

r\JTHIS DOCUMENT HAS BEEN REPRODUCED EXACTLY AS RECEIVED FROM THE

tr'N PERSON OR ORGANIZATION ORIGINATING IT. POINTS OF VIEW OR OPINIONS

CD STATED DO NOT NECESSARILY REPRESENT OFFICIAL OFFICE OF EDUCATION

POSITION OR POLICY.

Li/FINAL REPORT

Project No. 81073Grant No. OEG-9-8-081073-0117(010)

THE EFFECTS OF INDIVIDUAL DIFFERENCES ON

OBSERVATIONAL LEARNING IN THE

ACQUISITION OF A TEACHING SKILL

Frederick J. McDonald and Mary Lou Koran

School of Education

Stanford University

Stanford, California

March 1969

The research reported herein was performed pursuant to agrant with the Office of Education, U.S. Department ofHealth, Education, and Welfare. Contractors undertakingsuch projects under Government sponsorship are encouragedto express freely their professional judgment in the con-duct of the project. Points of view or opinions stated donot, therefore, necessarily represent official Office ofEducation position or policy.

(09 U.S. DEPARTMENT OF HEALTH, EDUCATION, AND WELFARE

(74..Office of Education

ei) Bureau of Research

O0

Qs_

ACKNOWLEDGMENTS

We would like to express our appreciation to

the many people whose help facilitated this study.

We are especially indebted to Dr. Richard E.

Snow, Dr. John D. Krumboltz and Dr. Nathan Maccoby, who

gave generously of their time, effort and encouragement.

Each of these persons, by his effective contribution made

possible the successful conduct of this research. To

them we extend our gratitude even if it can only be ex-

pressed in so modest a form as this acknowledgment.

Frederick J. McDonald

Mary Lou Koran

ii

...

TABLE OF CONTENTS

Page

ACKNOWLEDGMENTS ii

LIST OF TABLES vi

LIST OF FIGURES

SUMMARY

Chapter

I. THE PROBLEM

ix

xi

Definition of the Problem 1

Related Theory and Research 4

Observational Learning 4

Learning and Individual Differences . 9

Implications of Research onObservational Learning andIndividual Differences

Definition of Relevant AbilitiesAbility MeasuresThe Dependent VariableStatement of Hypotheses

1417222730

II. EXPERIMENTAL DESIGN 33

Subjects 33

Ability Measurement 34

Treatment, Procedures 38

Procedures for Model Preparation 42

Lesson Material 43Procedures for Training Microstudents 45

Criterion Measures 47Classroom Performance Measures 47Written Measures 51

Methods of Gathering Data 52Procedures for Training Raters 53

Rating Procedures 54

Rater Reliability 54

iii

7.,74T.MrMri,,,,V7M17,,,MT1,4711+71T,..7.,-","

Page

III. RESULTS 57

Instructional Treatment Main Effects 59Classroom Performance Measures 59Written Measures 79

Ability Measures 82Reliability 82Intercorrelations of Ability Measures . 84Relation of Ability Measures to

Performance 87Aptitude x Treatment Interactions . 92

Classroom Performance Measures 94Written Measures 105

IV. DISCUSSION OF RESULTS AND IMPLICATIONSFOR FUTURE RESEARCH 113

Summary of Data: Hypotheses Tests . . . 1.13

Instructional Treatment Main Effects 114Aptitude x Treatment Interactions 119

Implications for Future Research 135Implications for Educational Practice 139Limitations and Further Statistical

Considerations 141

143APPENDICES

A. Raters Manual 144

B . Set Induction Materials . 154

157

D . Written Measures 162

E . Rating Forms 166

F. Intercorrelations Among Dependent

C. Written Model

Variables 169

Analyses For Homogeneity Assumptions . . 173

H. Analyses of Variance: ClassroomPerformance Measures 176

iv

I. Newman-Keuls Tables: ClassroomPage

Performance Measures 185

J. Newman-Keuls Tables: Written Measures 194

K. Intercorrelations of Ability Measures 197

BIBLIOGRAPHY 201

LIST OF TABLES

Table Page

1. Organization of Task and Ability Variablesin a Multi-Process Model of Learning . 19

2. Number of Sample Subjects by Sex andSubject Matter 34

3. Ability Measures 37

4. Treatment Procedures 41

5. Rater Reliability II 56

6. Independent and Dependent Measures 58

7. Intercorrelations Among DependentVariables for Total Sample 60

8. Means and Standard Deviations ofDependent Variables 62

9. Analysis of Variance of Analytic Questions 64

10. Summary of Analysis of Variance ofAnalytic Questions: Simple MainEffects 65

11. Analysis of Variance of Categories ofAnalytic Questions 68

12. Summary of Analysis of Variance ofCategories of Analytic Questions:Simple Main Effects 69

13. Analysis of Variance of High QualityAnalytic Questions 72

14. Summary of Analysis of Variance of HighQuality Analytic Questions: SimpleMain Effects

vi

73

Table Page

15. Analysis of Variance of Nonanalytic76Questions

16. Summary of Analysis of Variance ofNonanalytic Questions: Simple MainEffects 77

17. Analysis of Variance of Scores forTrue-False Test 81

18. Analysis of Variance of Scores forMatching Test 82

19. Means and Standard Deviations ofAbility Measures 83

20. Reliabilities and Standard Errors ofMeasurement for Ability Measures 85

21. Intercorrelations of Ability Measures:Total Sample 86

22. Correlations of Ability Measures withPerformance Measures: Film-MediatedModeling Group . . o 89

23. Correlations of Ability Measures withPerformance Measures: WrittenModeling Group 90

24. Correlations of Ability Measures withPerformance Measures: Control Group . 91

25. Simple Regression Analyses forAnalytic Questions 95

26. Simple Regression Analyses for Categoriesof Analytic Questions 98

27. Simple Regression Analyses for HighQuality Analytic Questions . . 101

28. Simple Regression Analyses:True-False Test 106

vii

Table

29. Simple Regression Analyses:Matching Test

viii

Page

110

LIST OF FIGURES

Figure Page

1. A Multi-Process Model of Learning 17

2. Mean Frequency of Analytic QuestionsAcross Teaching Sessions 66

3. Mean Frequency of Categories of AnalyticQuestions Across Teaching Sessions 70

4. Mean Frequency of High Quality AnalyticQuestions Across Teaching Sessions .. 74

5. Mean Frequency of Nonanalytic QuestionsAcross Teaching Sessions 78

6. Disordinal Interaction 93

7. Interaction of Hidden Figures--I Scoreswith Instructional Treatments forCategories of Analytic Questions 96

8. Interaction of Hidden Figures--I Scoreswith Instructional Treatments forHigh Quality Analytic Questions . . 99

9. Interaction of Hidden Figures--I Scoreswith Instructional Treatments forHigh Quality Analytic Questions

10. Interacticn of Film Memory Scores withInstructional Treatments for HighQuality Analytic Questions

11. Interaction of Film Memory Scores withInstructional Treatments for True-False Test Scores

12. Interaction of Maze Tracing--II Scoreswith Instructional Treatments forTrue-False Test Scores

13. Interaction of Maze Tracing--II Scoreswith Instructional Treatments forMatching Test Scores

ix

102

103

107

108

111

Figure Page

14. Organization of Ability Variables in a Multi-Process Model of Learning 122

7111,1,71,

SUMMARY

The purpose of this study was to examihe the ef-

fects of verbal and perceptual dimensions of individual

differences in relation to the efficacy of two different

kinds of modeling procedures in the acquisition of a teach-

ing skill. The two different methods of presentation em-

ployed were written and film-mediated models. Both model-

ing treatments were expected to produce greater behavior

change than a control treatment. It was anticipated that

for Ss receiving the film-mediated model, criterion scores

should show stronger relation to perceptual abilities, while

for Ss receiving the written model, criterion scores should

show stronger relation to verbal abilities. These hypothe-

ses imply that there are nonparallel regression slopes, and

consequently, that one treatment will not be superior

throughout the distribution of perceptual and verbal abil-

ity variables.

Specific predictions, though tentative, were

based on theoretical considerations which suggested that

the requirements of the Written and Film-Mediated Modeling

treatments were sufficiently different to produce different

xi

ability-performance relationship. These predictions were

derived from an analysis of learning tasks and processes

corresponding to a theoretical model proposed for the in-

vestigation of individual differences in learning.

Following the administration of aptitude tests

selec -ted- from the French Kit of Reference Tests for Cogni-

tive Factors and from a series of specially developed audio-

visual testing materials, 121 Stanford intern teachers were

randomly assigned to three treatment groups: a Film-

Mediated Modeling treatment, consisting of a filmed por-

trayal of Analytic Questioning, the teaching skill to be

learned; a Written Modeling treatment consisting of a ver-

batim text of the sound track from the film-mediated model;

and a Control treatment which received no model, but went

through all other steps common to the two modeling treat-

ments;

Treatment conditions were held constant in all

ways except for the mode of model presentation. In terms

of general procedures all Ss received the initial instruc-

tions and microteaching pretest followed by two cycles com-

prised of models, rehearsal and microteaching Treatments

were terminated with the completion of two written tests.

The criterion performances assessed included the frequency,

xii

variety) and quality of Analytic Questions used upon three

separate teaching sessions in addition to scores on the

written measures.

Instructional treatment main effects as well as ap-

titude X treatment interactions were investigated. Analy-

sis of variance was used to test treatment effects. These

analyses disclosed highly significant treatment effects in

which Ss in both Written and Film-Mediated Modeling treat-

ments generated significantly higher frequency, variety and

quality of Analytic Questioning than did Control group Ss.

Similarly, Ss in the Written and Film-Mediated Modeling

treatments performed significantly better on both written

measures than did Control group Ss. Moreover, from the

average data alone, training under Film-Mediated Modeling

conditions appears to have been consistently more effective

than training under Written Modeling conditions across all

measures of the dependent variable.

Aptitude x treatment interactions were evaluated

by comparing regression slopes for different treatments.

Analyses of interactions disclosed that scores on Hidden

Figures, Maze Tracing and Film Memory tests interacted sig-

nificantly with the instructional treatments. In addition,

the magnitude of these interactions tended to increase

across performance trials. Interpretations of these inter-

actions were derived within the framework of the theoreti-

cal model described.

These findings suggest that the rate and level

of learning of a specific teaching strategy varies as a

function of model presentation; and that the effectiveness

of instructional methods varies from S to S with such dif-

ferences being related to trainee aptitudes. Results such

as these, if replicated, may provide a basis for the indi-

vidualization of teacher training programs.

xiv

CHAPTER I

THE PROBLEM

Definition of the Problem

The fields of education and psychology

abound with questions related to the attainment of

learning proficiency through variations in materials

and tasks, instructional techniques, and abilities

of the individual learner. There are also the more

complicated questions of the degree to which learning

effectiveness is a function of the interaction of

these variables.

Investigations of learning under different

instructional procedures commonly assign Ss randomly to

multiple treatment conditions and compare average perfor-

mance on some criterion measure. The usual conclusion

has been that one of the treatments is more effective

1

2

than others in some general sense. However, the meaning

of main effects is altered in the presence of interactions,

and variables not represented have no opportunity to dem-

onstrate their interactive effects (Snow, Tiffin, and

Seibert, 1965).

Investigators have often looked for aptitude

variables to serve as general predictors of criterion

scores, but this tends to focus attention on aptitudes

that correlate with outcome under almost any treatment

rather than on those likely to be involved in aptitude X

treatment interactions. Predictors having low or negative

correlations with success under certain treatments have

been largely disregarded, even though these variables may

be of considerable importance when instruction is adapted

to the individual (Cronbach, 1965).

As Cronbach (1957) has argued:

Applied psychologists should deal with treatmentsand persons simultaneously. Treatments are charac-terized by many dimensions; so are persons. The twosets of dimensions together determine a payoff surface(p. 680) . . . Ultimately we should design treatmentsnot to fit the average person, but to fit groups ofstudents with particular aptitude patterns whichcorrespond to (interact with) modifiable aspects ofthe treatment (p. 681).

Gagn4 (1964) also has suggested that individual

differences in the learner are among the most important

independent variables as one proceeds to the more complex

forms of learning. To this extent he was supporting the

Cronbach argument for simultaneous consideration of aptitude

and instructional variables.

One complex form of learning in which the role

of individual differences might be studied is that of ob-

servational learning. The wealth of literature surrounding

observational learning (Bandura and Walters, 1963; McDonald

and Allen, 1967) indicates the educational significance of

studying learning efficiency in association with modeling

procedures. Research in this area has consistently shown

that complex social responses and teaching skills may be ac-

quired or the characteristics of response hierarchies may be

considerably modified as a function of observing the behav-

ior of others and the consequences of their responses, with-

out the observer performing an overt response himself, or

receiving any direct reinforcement during the acquisition.

While Bandura and Walters (1963) have suggested

that observer characteristics influence the extent to which

observational learning occurs, observer characteristics

have been largely disregarded in previous research. Thus

a study of the relationship between individual differences

and the efficacy of different modeling procedures, while

4

exploratory, contains elements of potential importance

for evaluating the factors influencing learning efficiency

when modeling procedures are used in teacher training

programs.

Related Theory and Research

Observational. Learning

A common approach to the transmission or modifi-

cation of teaching skills has been to provide teacher

trainees with some type of written or oral instruction

followed by periodic feedback on their attempts to produce

these behaviors in the classroom. An alternative to this

strategy of teacher training is suggested by recent find-

ings on the role of observational learning in personality

development and behavior modification. Bandura and Walters

(1963), in a review of the relevant literature point out

that complex behavior may be acquired almost entirely

through imitation. The provision of live or symbolic

models serves to accelerate the learning process, and,

in cases where time or error is costly, this technique

becomes an efficient means of transmitting behavior pat-

terns (Bandura and Walters, 1963; Bandura, Ross and Ross,

1963). Other research has shown that the level of the

desired behavior exhibited by the observer can be as great

as that exhibited by the model (McBrearty, Marston, and

Kanfer, 1961). The implication of these findings for

teacher training is that the provision of live or symbolic

models displaying desired teacher behaviors may provide

an effective alternative to purely descriptive techniques

of training, (McDonald and Allen, 1967).

The concept of imitation in psychological theory

has a long history dating back to Lloyd Morgan (1896),

Tarde (1903) and the doctrine of instinct. In more recent

formulations it has been widely assumed that the Occurrence

of imitation, or observational learning, is contingent

upon the administration of reinforcement either to the

model or to the observer.

According to the theory proposed by Miller and

Dollard (1941), the conditions necessary for learning in-

clude a motivated subject who is positively reinforced

for matching the rewarded responses of a model.

Mowrer's (1960) proprioceptive feedback theory

similarly highlights the role of reinforcement, focusing

on the classical conditioning of positive and negative

emotions to matching response correlated stimuli. According

6

to Mowrer, imitative learning can occur either through

pairing of the model's response with reinforcers dispensed

to the observer, or by means of vicarious conditioning.

However, these reinforcement theories do not

account for the learning of matching responses when the

subject does not perform the model's responses during

the process of acquisition and for which reinforcers are

given neither to the model nor to the observer. The ac-

quisition of imitative responses under the latter condi-

tions can best be explained in terms of contiguity theory

of observational learning (Bandura, 1965; Sheffield, 1961).

According to this theory, it is assumed that when an ob-

server witnesses a model exhibiting a sequence of responses,

the subject acquires through the contiguous association

of sensory events, symbolic or representational responses

possessing cue properties capable of eliciting at a later

time, overt responses matching those that have been pre-

viously observed. Accordingly, Bandura (1963, 1965) has

argued that contiguity accounts for the acquisition of

matching responses, whereas reinforcement influences the

performance of imitatively learned responses.

It cannot be assumed that observational learning

is contingent solely upon exposure of an observer to a

7

complex sequence of stimulation. Bandura (1965) states

that imitation is an active process in which modeling

stimuli combine with other variables in shaping response

patterns. Factors other than contiguity undoubtedly in-

fluencing imitative response acquisition would include:

attention directing variables such as motivation, prior

training in observation, and anticipation of reinforcement;

rate, amount and complexity of stimuli presented to the

observer; practice; observer charanteristics and other

variables.

McDonald and Allen (1967) have applied the re-

search on observational learning to the learning of teach-

ing behavior, and have obtained data pertaining to the

following independent variables: (1) self-feedback and

reinforcement; (2) variations in feedback and practice

conditions; (3) film-mediated and written modes of model

presentation. These studies have indicated that reinforce-

ment and discrimination training administered by an ex-

perimenter were highly effective methods of producing

behavior changes in teachers. Because of considerable

information processing capabilities, human learners nay

require little in the way of feedback under many circum-

stances. However, if the responses to be learned are

8

sufficiently complex, some type of feedback may be required.

Moreover, exposure of an observer to a complex sequence

of teaching behavior does not guarantee that Ss will attend

to the most relevant stimuli, or accurately perceive the

cues to which their attention is directed. Discrimination

training, in which attempts have been made to focus Ss

attention on relevant stimuli, has been shown to facilitate

observational learning.

In the preceding experiments, it was observed

that different kinds of training procedures appear to be

differentially effective with trainees. While a great deal

of individual variation has been observed within the written

and film-mediated modeling conditions, differences between

the two modes of model presentation have not been consis-

tently significant. A promising line of research suggested

by these results would appear to be a study of the inter-

actions between the kinds of treatment used to produce

behavior change and (1) the type of teaching skill to be

learned, and (2) individual differences in the learner.

A study has recently been undertaken to investigate inter-

actions between various modeling procedures and types of

dependent variables (McDonald and Allen, 1967). The pur-

pose of the present study was to investigate the interaction

9

of modeling procedures with individual differences in the

learner.

Learning and IndividualDifferences

The relationship of tested abilities to learning

rate has been of experimental interest for quite some

time. Early investigators have examined the generality

of learning ability to determine if a single learning

ability accounts for performance in all learning situations.

Over a wide variety of learning tasks, it has generally

been found thr.t measures of learning in such tasks exhibit

low correlations both with other tasks and with intelli-

gence measures. The usual conclusion has been that there

is no general learning ability.

Husband (1939, 1941) was among the first to

conclude that learning ability is not unitary but that

there are a number of relatively independent learning

abilities depending upon the type of task. The most sub-

stantial evidence on this point, however, comes from a

series of studies by Woodrow (1938, 1939, 1940, 1946) in

which a group of students were given a number of practice

trials on a variety of learning tasks. A battery of special

10

aptitude and intelligence tests was administered both before

and after practice. The results obtained showed that there

was no single general learning ability accounting for im-

provement on all tasks. Moreover, there was no significant

relationship between scores representing gain with practice

and intelligence. Additional research in this area (Simrall,

1947) has indicated that even when the types of material

used in the learning tasks were highly similar to types

of material used in the intelligence tests, gains during

practice were still not significantly related to test

scores. While methodological weakness and inappropriate

statistical analysis (Rapier, 1962, Manning and Dubois,

1962) limit the usefulness of these early studies as a

basis for conclusions, later interpretations (Humphreys,

1960; Porter, 1959) have generally supported the position

that neither the amount gained nor the rate at which it

is acquired show any consistent relationship to intelligence

measures.

Related to the Woodrow trend is the work of

Fleishman and his associates (summarized in Fleishman,

1965) on the relationship of ability factors to performance

in the course of perceptual-motor skill learning. In-

vestigators of learning have commonly used a terminal

11

performance score as a criterion for learning. Such a

score is believed to be an accurate representation of the

learning that has taken place. However, Fleishman's find-

ings have suggested that the learning of complex skills

is a multidimensional process in which the contributions

of individual differences associated with performance on

a task change from early to later stages of learning.

These findings serve to underscore the importance of ex-

perimental work on behavioral change which investigates

the effects of task variables, particularly as they change

over conditions of practice or instruction, (Glaser, 1967).

The line of work generated by Woodrow is reflected

today in the work of other differential psychologists who

have related aptitudes to performance on various laboratory

tasks (Allison, 1960; Stake, 1961;Duncanson, 1964). Gen-

erally, in these studies, a battery of reference tests is

administered, the subjects complete various learning tasks,

and regression or factor analytic techniques are used to

relate learning to the reference tests. However, this

line of investigation has been primarily concerned with

task variables rather than instructional variables. While

these studies have investigated the relationship of apti-

tudes to performance under conditions of practice, they

12

have not generally compared ability-performance relation-

ships under different methods of instruction. Thus, these

results typically do not indicate any specific way in which

Ss could be treated or situations modified in order to

maximize learning.

Direct pursuit of interactions between treatment

and learner variables is infrequently represented in the

literature. While interactions with instructional variables

have most commonly been reported.as incidental findings

in experiments designed for other purposes, there are a

few studies that have deliberately investigated relation-

ships among learner characteristics and learning outcomes

under different instructional conditions.

An investigation of individual differences

and instructional film effects (Snow, Tiffin and Seibert

1965), using filmed and live lecture demonstrations found

that a number of variables including ascendancy, responsi-

bility, numerical aptitude, verbal aptitude, past experience

with entertainment films, and past use of college library

instructional films interacted significantly with instruc-

tional treatments.

Edgerton (1958), using Navy training courses,

taught the same material by two different methods in an

13

effort to modify the conventional rote instruction into

a meaningful method. Correlations obtained indicated that

the Thurstme memory and fluency factors interacted with

the instructional conditions.

Grimes and Allinsmith (1961) compared primary

reading achievement under two methods; a structured phonics

program, and a less structured whole-word approach, with

anxiety and compulsivity used as differential variables.

It was anticipated that highly anxious or highly compulsive

children would show more achievement under structured

methods than similar children taught by the unstructured

method. This expectation was supported. Moreover, anxiety

and compulsivity, which were not correlated in their sample,

interacted with each other as well as with instructional

method.

Finally, Cronbach (1965), in a partial survey

of the literature dealing with attitudes concerning con-

fidence, willingness to risk failure, and motivation for

self directed achievement, examined numerous indications

from the work of Atkinson (1964), Wallach and Kogan (1964)

and others that certain personality variables interact

with instructional variables.

It seems clear that interactions may be dis-

covered in a random exploration in which results under

14

multiple instructional treatments are projected onto a

mass of differential information. Cronbach (1965), however,

argues for what he considers a more promising experimental

strategy in which alternative treatments are designed to

interact with specific differential variables. The design

of the present study corresponds more closely to the latter

strategy in that treatments were initially selected in

terms of specific aptitude variables and then refined to

maximize the effects of these variables.

Implications of Research onObservational Learning andIndividual Differences

What are the implications of the previous dis-

cussion for study of the interaction of individual differ-

ences with modeling procedures in the acquisition of a

teaching skill?

First of all, instructional methods differ. The

literature reviewed here suggests that a person learns

more easily from one method than another, that this best

method differs from S to S and that such differences be-

tween treatments are related to learner characteristics.

Different methods of presentation may be employed

in the use of modeling procedures. That is to say that

15

different methods of presentation may be used to reach

the same terminal objective. Among these are written and

film-mediated models. For the purposes of this study,

by "exposure to a film-mediated model" is meant that the

learner has observed the actual filmed performance of

another person who displayed the behaviors to be acquired.

By "exposure to a written model" is meant that the learner

has read a written transcription of the sound track from

the film-mediated model. While both of these methods

have been effective as training procedures, differences

between the two treatments have not consistently been

found (McDonald and Allen, 1967; Bandura and Mischel,

1965). However, in view of the task differences generated

by the two different modeling treatments, it seems reason-

able to expect that different abilities may also be in-

volved.

A study by Wolfgang Boehm (cited by Cronbach,

1957) tends to lend support to this expectation. In this

study, a sound film was shown to the experimental groups.

Matched control groups read a verbatim text of the sound-

track. Differences in comprehension between the two groups

were insignificant. However, a general mental test cor-

related only .30 with text learning, but predicted film

16

learning with an average correlation of .77. This dif-

ference was consistent across all ages and would appear

to support the hypothesis that different abilities might

predict success on these two different methods of presenta-

tion.

In answer to the question of why purely audio-

visual and purely written verbal treatments should be used

when a combination of the two might be likely to produce a

stronger training effect for all, or for any one or more

sub-groups, it is suggested that such a separation is a

useful research strategy which may offer a clearer under-

standing of the separate functions of the treatments and

abilities involved. A mixed treatment might mask the

functioning of aptitudes which may be interacting with

treatment effects. Moreover, the two treatments utilized

here have already been defined as meaningfully distinct by

previous research. These treatments have both theoretically

and practically distinguishable features. As Cronbach (1957)

has argued, unless one treatment is clearly best for every-

one, a situation which has not occurred for the treatments

under consideration here, treatments should be differen-

tiated in such a way as to maximize their interaction with

aptitude variables. In this way, if interactions do exist,

learning is maximized.

17

Definition of Relevant Abilities

While individual differences in performance on

learning tasks can be measured as a function of numerous

operationally defined variables in those tasks, and factor

analytic studies can relate performance on a number of

tasks to a variety of reference tests, ultimately, research

on individual differences must be guided by theories of

human learning and performance. As Melton (1967) has argued:

What is necessary is that we frame our hypothesesabout individual differences variables in terms of

process constructs of contemporary theories of learn-

ing and performance (p. 239) .

Accordingly, Melton has proposed an expanded

version of a theoretical model initially described and

tested by McGuire (1961) as a framework for investigating

individual differences in learning. The model is shown

in Figure 1.

Figure 1. A Multi-Process Model of Learning

18

This model proposes a stimulus differentiation

component, a response integration component, and a "hookup"

between the encoded stimulus and the integrated response

or a segment of it. The stimulus differentiation com-

ponent (S1 ri) is the subjects' coding response to

the potential physical stimulus (SO. It is the effec-

tive, or functional stimulus (ri(si)) in S-R associations,

and reflects such factors as learning and set as well

as the physical stimulus. The response integration com-

ponent (RaRbRc) is the output response, which may be

either a previously learned unit (Ra) or a new combination

of such units, (RaRbRc). The "hookup" is the connection

(ri(s1)---4wRaRbRe) between the functional stimulus and

the required response. The component (rm(dm)') represents

an alternative mediational route for the connection of

the internal representation of the physical stimulus and

the required response.

Differences and similarities in task and ability

variables involved in the two modeling procedures previously

described may be represented in terms of this paradigm.

This representation is summarized in Table 1.

First, Ss in the Film-Mediated Modeling condition

view an actual portrayal of the teaching skill and of the

AIN

19

TABLE I

ORGANIZATION OF TASK AND ABILITY VARIABLES

IN A MULTI-PROCESS MODEL OF LEARNING

S1 _IN, ri( si) R aRb Rc

StimulusDifferentiation

AssociationMediation

ResponseIntegration

WrittenModel

Read script of soundtrack

Generate associativecontext

Set own reading pace Transform verbal tobehavioral repre-sentation

Reread, pause asneeded

Retain representation

Generate new associates

Integrate in behavioralcontext

Code as verbalrepresentation

View film, hearsound track

Generate associativecontext

Retain representations

Film-MediatedModel

Keep pace with film Transform behavioralto verbal repre-sentation

Generate new associates

Review in memory Integrate in behavioralcontext

Flexibility ofclosure

Verbal BehavioralTransformation

Verbal memory

Ability Perceptual speed Verbal Association Audio-Visual memory

VariablesSpatial scanning Verbal fluency

Short-term memory

Verbal comprehension

20

pace of the lesson. This involves processing information

from multiple channels simultaneously. Film-mediated

models are extremely rich in perceptual detail in that

there are many different cues to which to attend, including

many extraneous to the relevant features of the model.

These task characteristics would appear to require the

ability to keep definite task-relevant dimensions in mind

so as to make identification in spite of perceptual dis-

traction. This ability may be similar to that commonly

called flexibility of closure (Thurstone, 1944).

Moreover, Ss in the Film-Mediated Modeling condi-

tion are expected to perceive and encode events as be-

havioral representation at the speed of presentation of

the film. Consequently, they are able to review relevant

material in memory only. The requirements of a predeter-

mined pace would seem likely to involve speed in perceptual

evaluation and in exploring a complicated spatial field,

as well as short term memory facility.

In the Written Modeling condition, Ss read a

written transcription of the soundtrack from the film-

mediated model. A written transcript of the model's be-

havior, however, presents only verbal components of these

stimuli. Consequently, Ss in the Written Modeling condition

23.

process information from a single channel in which they

are able to establish their own pace and return to review

relevant material as read. While Ss in the Written Modeling

condition must also distinguish the significant features

of the model, the total set of stimuli to select from is

limited to verbal components. These task variables would

be expected to require the ability to comprehend written

verbal material.

Corresponding to the associational-mediational

component of the model, it is hypothesized that Ss in both

Film-Mediated and Written Modeling conditions must generate

their own analogy to the model's performance. While Ss

in the Film-Mediated Modeling condition are expected to

abstract relevant verbal representation from behavioral

representation, Ss in the Written Modeling condition are

expected to generate relevant behavioral representation

from solely verbal components. These task variables would

appear to require facility in both verbal association

and in verbal-behavioral transformation. Although these

abilities would appear to be involved in both treatments,

they are presumed to be more decisively involved in the

Written Modeling condition. Because of the abstraction

of the script, or compression of stimuli into solely verbal

22

components, a greater degree of verbal processing would

appear necessary for those in the Written Modeling

condition to generate associative context and behavioral

representation. Ss in the Film-Mediated Modeling condition

have only to recall much of what Ss in the Written Modeling

condition must produce.

Finally, corresponding to the response integration

component of the model, it is expected that Ss in both

conditions need to retain representations of their respec-

tive models, generate new examples of the skill being

modeled in the teaching situation, and integrate these

examples into a behavioral context. These task variables

would be expected to require both verbal fluency, and the

ability to remember major ideas as well as details from

audio-visual presentations in the case of Ss in the Film-

Mediated Modeling condition, and verbal presentations in

the case of Ss in the Written Modeling condition.

Ability Measures

The selection of ability measures for use in this

study was made from the Kit of Reference Tests for Cognitive

Factors (French, Ekstrom and Price, 1963) and from a series

of film and audio tests developed by Seibert and Snow (1965)

23

and Seibert, Reid and Snow (1967). The abilities assessed

are believed to be those which distinguish most clearly

between the two modeling conditions and are consistent

with the analysis of the Film-Mediated and Written-Modeling

conditions previously discussed.

The Kit of Reference Tests for Cognitive Factors

was developed from a project organized to select tests to

represent each of the better established factors in the

cognitive behavioral domain. It consists of a group of

tests representing each of the more frequently obtained

factors in the cognitive ability area, rather than a stan-

dardized battery of tests. The purpose of the kit is to

provide research workers with a set of tests for defining

each of these factors and for facilitating interpretation

and comparison of one factor study with another.

The manual does not provide the reliability,

forming, validity or other information usually included

in a test manual. Such information has not been included

because these tests are suggested for the single purpose

of factorial research. Consequently, there is no compre-

hensive information on how these tests intercorrelate.

Where there is information on some of the tests, it has

been gathered on samples dissimilar to the one used in

24

this study. However, all of the tests have two separately

timed parallel parts to permit the estimation of reliability,

and a correlation matrix has been computed to provide

additional necessary information for subsequent discussion.

The film and audio tests were developed in an

effort to examine alternatives to conventional printed

paper tests in psychometric research. It is argued by

Seibert and Snow (1965) that since human beings behave

in contexts which are not static, not primarily verbal,

and not characterized by the presence of print, it is

conceivable that some of this behavior involves abilities

or predispositions which are not amenable to adequate

sampling with printed paper tests.

The purpose of their investigation was to begin

the exploration of individual differences in performance

on tasks characterized by moving, sequential or behavioral

content. Since those characteristics were involved in

the present study, it was thought that these tests would

provide useful information. Tests chosen from the Kit of

Reference Factors for use in this study have also appeared

prominently in the factor interpretations of the earlier

film test study.

25

A brief description of the ability tests and

the factors they represent follows:

Hidden Figures (Cf-1, Flexibility of Closure).

A test of the ability to keep one or more definite con-

figurations in mind so as to make identifications in spite

of perceptual distraction. The visual and auditory modali-

ties of the closure factor have been shown to be highly

correlated. (White, 1954)

Identical Pictures (P-3, Perceptual Speed).

A test of speed in finding figures, making comparisons

and carrying out other tasks involving rapid checking

and perceptual evaluation.

Maze Tracing (Ss-1, Spatial Scanning). Related

to perceptual speed is spatial scanning ability, which

is described as speed in visually exploring a wide or

complicated spatial field. Tests of perceptual speed

and spatial scanning tended to appear together on a film

test factor (Seibert and Snow, 1965) hypothesized to in-

volve rapid serial processing of spatial images, and sug-

gesting the dynamic information processing nature of this

ability.

26

Advanced Vocabulary (V-4, Verbal Comprehension).

A test of the ability to understand the English language.

Individual differences are most clearly seen in the size

of comprehension vocabularies, but also exist with respect

to tests demanding knowledge and understanding of grammati-

cal patterns, sentences, phrases and other aspects of the

English language.

Word Arrangement (Fe-3, Expressional Fluency).

A test of ability to think rapidly of appropriate wording

for ideas. The emphasis in this test is on facility in

producing connected discourse that will fit restrictions

imposed in terms of given words.

Film Memory. A test of the ability to iden-

tify verbal descriptions from remembered behavioral con-

tent. It consists of a motion picture presentation,

followed by printed questions concerning the action por-

trayed.

Memory for Ideas. A test of ability to ab-

stract major ideas from remembered audio taped verbal

presentation.

27

Sentence Reproduction. A test of the ability

to reproduce verbal material in detail from remembered

audio taped verbal presentation.

GRE Verbal Aptitude scores were also used in

this study. This score represents verbal reasoning and

reading comprehension.

The Dependent Variable

Educators hold that one of the major objectives

of instruction in the schools is the development of skills

or strategies which will enable students to become life-

long autonomous learners. While the content of any field

may undergo rapid change, the ability to analyze crit-

ically written materials within that content area is

an objective of any field of study (Bloom, 1956). One

way of developing such ability in students is to train

teachers to ask the kinds of questions that require stu-

dents to engage in the analysis of their reading material.

With this objective in mind, the dependent variable de-

veloped for use in this study is termed Analytic Ques-

tioning.

The basic ideas underlying the development of Ana-

lytic Questioning_ were derived from Taxonomy of Educational

28

Objectives (Bloom) 1956) which presents a plan for classi-

fying educational objectives. This taxonomy was developed

from a project organized to achieve a more widely accepted

set of behavioral objectives that could be more profitably

used for communication and for guiding research in connec-

tion with curricula, teaching and examining in education.

Bloom and his associates claim that any educa-

tional objective can be classified within their taxonomy,

and imply that any question can also be classified. One

of the ways in which each category was defined was by using

examples of questions that required students to engage in

specified kinds of behavior (Sanders, 1966). Accordingly,

if the objectives of the taxonomy are considered to be

worthwhile educational objectives, training teachers to

formulate questions requiring students to engage in the

behavior specified as an objective would consequently be

considered a worthwhile training objective for prospective

teachers. Thus, analysis of elements, a subset of the

general objective of Analysis, was selected as an appropri-

ate source for training teachers in questioning behavior.

The Taxonomy of Educational Objectives defines

Analysis in the following manner:

29

The breakdown of a communication into its constituentelements or parts such that the relative hierarchy ofideas is made clear and/or the relations between theideas expressed are made explicit. Such analyses areintended to clarify the communication, to indicate howthe communication is organized, and the way in whichit manages to convey its effects, as well as its basisand arrangement (p. 205)

Although Analysis may be conducted simply as an

exercise in identifying structure and organization in a

communication, it is probably more educationally useful

as an aid to fuller comprehension or as a prerequisite

to higher level skills in the taxonomy, such as Synthesis

or Evaluation.

The Analysis of Elements is specifically con-

cerned with the breakdown of a communication into its

constituent parts, i.e. to identify or classify the elements

of a communication. Any communication may be conceived

of as composed of a number of elements, some of which are

e xplicitly stated in the communication and can be recognized

and classified relatively easily. However, there are other

e lements in a communication which are not as clearly labeled.

These elements can be inferred only from an analysis of

a series of statements within the communication. Until

the reader can detect them, he may have difficulty in fully

comprehending or evaluating the communication.

30

A detailed discussion of the dependent variable

appears in the Raters Manual (Appendix A). Thus comments

here will be brief. The categories of Analytic Questioning

are the following:

1. Identification of Hypotheses

2. Semantic Definition

3. Identification of Unstated Assumptions

4. Distinction of Factual from Non-Factual

Statements

5. Identification of Conclusions

The labels in each case generally reflect the teacher's

goal when using a given category of questions.

Statement of Hypotheses

Based upon the previously reviewed research

and theory, the basic hypotheses derived and tested were:

1. Both Film-Mediated and Written Modeling

conditions will produce significantly greater

changes in the response strength of desired

behaviors than will the Control condition

2. For subjects receiving the Film-Mediated

Modeling condition, criterion scores should

show stronger relation to perceptual abilities

31

thEn for subjects receiving the Written

Modeling condition.

3. For subjects receiving the Written Modeling

condition, criterion scores should show

stronger relation to verbal abilities than

for subjects receiving the Film-Mediated

Modeling condition.

These hypotheses imply that there are nonparallel

regression slopes, and consequently, that one treatment

will not be superior throughout the distributions of the

perceptual and verbal ability variables. For the purpose

of this study the term "perceptual abilities" has been

used to include both tests of figural cognition from the

French Kit and the audio visual memory tests. While the

audio-visual tests are experimental in nature, and it is

not yet clear precisely how they should be classified,

they have been tentatively grouped with the tests of fig-

ural cognition, mainly because of their mode of presenta-

tion. The term "verbal abilities" has been used to include

written tests of verbal cognition.

Although perceptual and verbal abilities are

referred to here in a seemingly unitary sense, it should

be emphasized that tests representing these abilities

32

are presumed to be independent factor tests. While it

may be anticipated that both perceptual and verbal abili-

ties would display a tendency to correlate more highly

with one another than with other more dissimilar measures,

identical ability-performance relationships have not been

anticipated for all tests of a given class. The relation-

ship of any particular test labeled "perceptual" or "verbal"

to any particular instructional outcome may be expected

to vary as a function of its relevance to both task and

process variables.

CHAPTER II

EXPERIMENTAL DESIGN

Subjects

Intern subjects were drawn from the Stanford

Teacher Education Program. They were initially categorized

by the subject matter they were teaching, and subsequently,

randomly assigned to one of three treatment conditions.

All of the subject matter areas of the Stanford Teacher

Education Program were represented with the exceptions of

foreign language and physical education majors. In both

of these cases it was felt that the prescribed methodology

was inappropriate for the systematic use of the dependent

variable, Analytic Questioning. Moreover, use of the audio

tapes of foreign language majors would have created undue

transcription and measurement problems.

Relevant characteristics of the sample studied

appear in Table 2 which provides the distribution of sub-

jects by sex and subject matter.

The Ss ranged in age from 21 to 40 years, al-

though the variability was relatively small with most of

33

TABLE 2

NUMBER OF SAMPLE SUBJECTS BY SEX AND SUBJECT MATTER

N = 121

Subject Matter Mai or

TreatmentGroup N

Sex

4m..-1

1-4

too

W

wr-i 0as o.0-1 WC) rl0 0ca ca

w0oW

4-10co

1 ti

cu 04-H+3 -I-)cd cdX M

M F

Film-MediatedModeling

WrittenModeling

Control

41

40

40

8

9

12

33

31

28

13

11

11

17

18

16

5

4

5

3

3

3

0..-1

+3k

m0

cg X

2 2

2 3

2 1

35

the ages falling near the mean of 23.0. The mean under-

graduate Grade Point Average of this group was 3.03.

Interns taught each of their lessons to a micro-

class of four students. These students were randonly as-

signed to teams of four which were subsequently randomly

assigned to Ss within treatment groups. The microstudents

were drawn from the Palo Alto school system and were paid

for their services. A total of 120 students were hired.

All students were either tenth or eleventh graders with

a grade point average of 3.0 or better. The majority of

these students were from middle class backgrounds.

Ability Measurement

Ss were tested prior to their participation in the

experimental procedures. Tests were administered in a

single two and one-half hour session on two separate oc-

casions. Testing time included instructions, practice

exercises and a fifteen minute rest interval.

In the first testing session, it was impossible

for all Ss to be tested together in the same room because

of the size of the intern group and the nature of the

36

audio-visual tests. Consequently, the one large group was

divided into two smaller groups which were tested simul-

taneously, in separate rooms. Moreover, the amount of

equipment involved in the administration of the audio-

visual tests, and the limitations imposed by physical

facilities did not permit the duplication of audio-vi;i,w1.

facilities in both rooms. Therefore, each group was ad-

ministered the tests in two separate sets--a set of audio-

visual tests and a set of paper and pencil tests from the

Kit of Reference Tests for Cognitive Factors. Upon com-

pleting the first set of tests, subjects changed rooms

and completed the second set. Within each of these two

sets the order of test administration was randomly assigned.

A second testing session was provided for those

Ss unable to attend the first session. This group was

small enough to permit all Ss to be tested together. Thus,

in this group it was possible to administer the two sets

of tests in a random mixed order. A summary of the testing

procedures appears in Table 3.

Tests were hand scored by three trained scorers

working under the direction of the experimenter. Objec-

tive test scoring keys were used on all tests with the

exceptions of Word Arrangement, Memory for Ideas and

TABLE 3

ABILITY MEASURES

*N.

NumberTest Time of

Parts

Presentation Order

Session One Session Two1 2 1

N=45 N=42 N=34

Kit of Reference Tests forCognitive Factors

Hidden Figures (Cf-1) 20 min 2 1 4 7

Identical Pictures (P-3) 3 min. 2 2 5 8

Maze Tracing (Ss-1) 6 min. 2 3 6 3

Word Arrangement (Fe-3) 8 min. 2 4 7 2

Advanced Vocabulary (V-4) 8 min. 2 5 8 4

45 min.

Audio-visual Tests

Film Memory 7 min. 1 6 1 1

Memory for Ideas 16 min. 2 7 2 5

Sentence Reproduction 34 min. 2 8 3 6

57 min.

38

Sentence Reproduction. Scoring directions for these tests

were sufficiently unambiguous as to cause little scoring

uncertainty. When such uncertainty did arise, it was

resolved by the experimenter. Scoring was checked for

accuracy. Both part scores and total scores were recorded

for each test. GRE-Verbal Aptitude scores were available

from Stanford Teacher Education Program records.

Treatment Procedures

A total of 121 Ss were used in this study. All

Ss had been administered the selected ability measures

prior to the presentation of the learning task. Each S

was randomly assigned to one of three treatment conditions:

a Film-Mediated Modeling condition in which subjects were

exposed to filmed portrayal of the behavior to be learned,

Analytic Questioning; a Written Modeling condition in

which Ss read a verbatim text of the sound track from the

film-mediated model; and a Control group which received no

model, but went through all other steps common to the

previously mentioned two groups.

The treatment procedure may be broken down into

nine steps. In seven of these steps, all groups received

39

identical treatment. In the remaining two, each group

was exposed to the appropriate type of modeling procedure.

Thus, treatments were held constant in all ways except

for the mode of model presentation.

The treatment procedure was broken down into

the following steps:

1. Set Induction--Ss were presented with brief writ-

ten instructions describing the nature of the

learning task (Appendix B).

2. Teaching Session One, (Ti)--A preevaluation of

each S was made prior to the presentation of the

model as each intern presented a ten minute les-

son to a microclass of four students. Criterion

scores from this lesson provided a baseline fre-

quency of the desired behavior.

3. Model Lesson One--Depending on their assignment

to an experimental condition, Ss were exposed to

a written or film-mediated version of a model

exhibiting Analytic Questioning behavior. The

Control group was given a set of written materials

extr= aneous to the learning task in order to fill

the equivalent time block,.(Appendix C).

40

4. Rehearsal One--Ss were given five minutes to in-

corporate the modeled behavior into their planned

lesson.

5. Teaching Session Two, (T2)--Ss taught a second

ten-minute lesson.

6. Model Lesson Two--Ss were exposed to the same

model presentation as used in step 3.

7. Rehearsal Two--Ss were given five minutes to in-

corporate the modeled behavior into their planned

lesson.

8. Teaching Session Three, (T3)--Ss taught a final

ten-minute lesson.

9. Testing Session--Ss were given two written tests

covering material presente,1 in the treatment pro-

cedures (Appendix D).

A summary of the procedures appears in Table 4.

All Ss thus received written instructions followed by two

cycles comprised of models, rehearsal, and teaching. Treat-

ments were terminated with the completion of the written

tests. Ss taught a different group of students in each

lesson. This permitted them to revise the same basic sub-

ject matter while attending to the improvement of Analytic

Questioning behavior.

TABLE 4

TREATMENT PROCEDURES

Steps GroupTime

Film-MediatedModeling

Written Modeling Control

1. Set Induction X X x 5 min.

2. Teach 1 X X X 10 min.

3. Model 1 film mediated written extraneousmaterial

10 min.

4. Rehearsal 1 X X X 5 min.

5. Teach 2 X X X 10 min.

6. Model 2 film mediated written extraneousmaterial

10 min.

7. Rehearsal 2 X X X 5 min.

8. Teach 3 X X x 10 min.

9. Test Administration X X x 5 min.

Explanation of Symbols: (X) indicates that all Ss received thisstep of the treatment in an identical manner. Written descriptionsare provided for the two steps in which treatments varied amongthe three groups.

42

Procedures for Model Preparation

An experienced teacher was selected to portray

Analytic Questioning in the film-mediated model. The

model was thoroughly familiarized with both the skill to

be modeled and the desired quality level through written

materials, discussion, and illustration of representative

examples from each category of Analytic Questioning. A

ten-minute lesson was planned by the model and experi-

menter together to provide at least two representative

examples of each category of Analytic Questioning. Micro-

students participating in the videotaping of the model

were provided training along with the teacher on the cate-

gories of questions to be asked, and appropriate responses

to each of these.

Several trials were provided for taping the

model lesson. During each trial, the experimenter recorded

the number and quality level of the Analytic Questions

used, and noted suggestions for improvement. The model

and experimenter would subsequently review the lesson

previously taught and make revisions where necessary.

From this pool the best of the tapes was selected.

Upon completion of the film-mediated model, a

brief commentary was superimposed on the soundtrac to aid

43

the discrimination of important behaviors in the lesson.

A verbatim text of the soundtrack was then typed from the

film-mediated model to provide the written model (Appen-

dix C) . The commentary superimposed on the soundtrack

of the film-mediated model appeared on the written model

in exactly the same form and place in the lesson.

Lesson Material

Because the extent to which Analytic QuestioniRE

can be systematically used is partially a function of the

opportunities provided for such questioning by the lesson

material at hand, lesson material was preselected for Ss

in this experiment. With the help of both faculty members

and intern supervisors, material was selected to meet the

following criteria:

1. The lesson material had to provide a rich source

for the systematic use of Analytic Questioning.

2. The lesson material had to be adaptable to the

ten-minute teaching interval established.

3. The content of the lesson material could not pre-

suppose a substantial amount of prior information

on the part of the microstudents.

44

A total set of thirteen written communications

was finally selected to serve as lesson material. The

number of lessons selected in each subject matter area

represented a ratio of approximately one lesson for every

eight interns in that subject matter area. This was done

to avoid overexposure of microstudents to any given lesson,

and also to demonstrate the utility of Analytic Question-

ing, as a questioning strategy, over a variety of materials.

The topics represented in this selection included student

activism, the hippie culture, the use of drugs, the defi-

nition of patriotism, the cultural value of pop music,

and others. Sources for this material included newspaper

and magazine articles in addition to material especially

written for the purposes of this study.

Lessons were randomly assigned to Ss within sub-

ject matter areas and treatments, and were distributed to

them one week prior to their participation in the experi-

ment. Interns were instructed to come prepared to teach

a ten-minute lesson discussing this written material with

the microstudents. In addition, they were told that the

microstudents would also have copies of this material which

they would be prepared to discuss at that time.

Procedures for Training Microstudents

45

Training was provided for microstudents for the

following reasons: (1) the complexity of the dependent

variable, (2) the ceiling effect on its frequency of oc-

currence imposed by time and availability of appropriate

student responses.

Treatments were designed to produce increases

in Analytic Questioning behavior. However, both the ques-

tioning strategy and the lesson material selected were

complex. While the ability level of the microstudents

was specified with this in mind, the questioning strategy

is tied to the analysis of the written material at hand.

Unless student responses were highly available, much of

the tenminute lesson period could have been spent in

searching the written material for appropriate answers,

limiting the frequency with which such questions could

be asked by the teacher. This would serve to mask the

functioning of treatment effects. Accordingly, optimal

conditions for the demonstration of treatment effects

would necessarily include high response availability on

the part of the students.

Training of the microstudents was accomplished

in the following manner. One week prior to the experiment,

46

all microstudents attended an orientation meeting in which

copies of each set of lesson materials to be used for

analysis were distributed along with a description of the

categories of questions which they could expect to be asked

with respect to that material. Examples of appropriate

questions and answers for each lesson corresponding to

each of the categories of Analytic Questioning were pro-

vided. Students were instructed to be well prepared on

this material as their participation in microteaching

would be contingent upon their successful completion of

a written test on the material to be administered prior

to each day of microteaching in which they were to par-

ticipate.

A random sample of four lessons was selected as

a source for this test prior to each day of microteaching.

Students were instructed to list at least one example of

material corresponding to each of the categories of Ana-

lytic Questioning for each lesson. These tests were checked

by the experimenter together with assistants. No micro-

student failed to meet these minimal criteria.

47

Criterion Measures

Consideration was given to several aspects of

Analytic Questioning skill. As a result both classroom

performance and written criterion measures were developed

for use in assessing treatment effects and ability-

performance relationships.

Classroom. Performance Measures

Several criterion measures were obtained from

transcripts of the classroom interaction of intern and

microstudents over three teaching sessions.

First, treatments were designed to produce in-

creases in Analytic Questioning as a general questioning

strategy. While different levels of significance might

be examined in relation to each subcategory of Analytic

Questioning_, differences of this nature were not emphasized

in the model presentation. Therefore, the total number of

analytic questions was selected as one appropriate index

for measuring increases in Analytic Questioning.

In addition to the total frequency, the variety

of categories of Analytic Questioning used was also of con-

siderable interest. Theoretically, the use of any category

of Analytic Questioning by Ss :represents an available class

of responses in the behavioral repertoire of that S as

well as a single occurrence of a behavior of interest.

Accordingly, for the purpose of this study, a total fre-

quency score representing the use of all categories would

demonstrate more desirable training effects than would

the same total frequency score representing the use of

fewer categories of Analytic Questioning.

Finally, while attention so far has been focused

exclusively on frequency measures, it was recognized that

the quality of the questions might also be considered a

potentially important dimension of the dependent variable.

Some analysis of quality had already been done in specify-

ing those kinds of statements fitting into the category

system. These categories were meant to include the best

examples of the behavior being learned. While weaker or

more ambiguous examples, which would not clearly meet the

criteria imposed by the category definition, would undoubt-

edly represent levels differing in quality, a considerable

degree of precision in classification was believed necessary

so that the extent of imitation could be assessed. However,

in developing the rating manual additional differences of

such a character were considered.

49

Quality is commonly defined as an approximation

to an ideal or standard. Accordingly, for the purposes of

this study, "quality" was used to distinguish between near

versus more remote approximations to the behavior exhibited

by the model. It had been observed in a previous study

(Koran, 1967) using a similar questioning strategy, that

one dimension along which questions in any given category

tended to vary was that of information content. Thus, in-

formation content was selected as the dimension along which

approximation to the behavior of the model would be assessed.

In all cases, the questioning behavior exhibited by the

model in this study required the student to supply the maxi-

mum amount of information that could be elicited by that

particular type of question. The closest approximations

to the model's behavior would also do this. In more remote

approximations, the teacher would supply varying amounts

of information which the model had elicited from the stu-

dents. Accordingly, questions meeting the criterion es-

tablished for high quality were designed to elicit from

the students all information relevant to answering them.

Questions requiring students to supply only part of the

information relevant to answering them, or requiring only

that students agree, disagree or select from among given

alternatives would not meet the criteria imposed by this

specification of quality.

To illustrate: the information content of a

semantic definition has been defined as a statement of the

methods and standards used for defining a word. The ques-

tion "How does the author define social class"? requires

students to supply the key elements of the definition as

well as the method of measurement, and would accordingly

be classified as a question of high quality. However, the

questions "How does the author define social class with

respect to income level"? and "Does the author define social

class in terms of occupation, education or income ? provide

part or all of the specified information content, and thus

would not meet the criteria established for high quality.

A complete discussion of quality, as used here, appears

in the Rater Manual ('ppendix A).

It should be noted that quality, as defined here,

is not intended to represent an evaluation of which kinds

of questions are "best." Any question could possibly be

made "better" or "worse" by changes which do not involve

approximation to the behavior of the model along the dimen-

sion of information, content. Attention is focused on the

quality dimension used here only to the extent that it can

be used to assess the extent of imitation.

51

Written Measures

While ultimately, decisions in teacher training

must be guided by the effectiveness of training procedures

in the transmission or modification of actual classroom

teaching behavior, written measures were introduced in an

effort to gather additional information useful in assessing

ability-performance relationships in observational learning?

These measures consisted of two recognition type tests

(Appendix D). These written tests were designed in the

following manner:

1. A true-false test in which subjects were asked

to identify the major categories of Analytic Ques-

tioning. (2 minutes)

2. A multiple choice test in which subjects were

asked to match a number of questions according

to membership in a particular subcategory of

Analytic Questioning. (3 minutes)

A common observation has been that differences

exist between measures of knowledge or comprehension, and

measures of actual performance in a given situation (Cron -

bath, 1963). In terms of the model proposed for investi-

gating individual differemces in learning that was pre-

viously discussed, ability-performance relationships may

52

be expected to vary as a function of the nature of the

task. While certain abilities, such as those associated

with verbal production, may be more decisively involved in

classroom performance measures, it is conceivable that dif-

ferent abilities, such as those related to perception and

memory, may be more likely to be involved in recognition

measures of the type used here.

Methods of Gathering Data

During the study each of the intern's lessons

was recorded on stenorette tape. Technicians collected

the tapes at the end of each lesson and labeled them with

each Ss n am e , treatment group and number of teaching ses-

sion. Each tape was monitored in order to assure that the

volume and length were adequate. Typists then prepared a

written transcript of each lesson. Each stenorette used

for transcribing was equipped with a timing device which

permitted the accurate recording of the ten minute com-

pletion time.

Written tests were collected following the third

teaching session. Each test was labeled with Ss name and

treatment group. Objective scoring keys were used for

both tests.

53

Procedures for Training Raters

Prior to the analysis of the experimental tapes,

a preliminary manual was developed from the written de-

scription of the dependent variable and analysis of the

model lesson. A team of four raters was given intensive

training on the definition of each category of Analytic

Questioning and the use of the preliminary rating manual.

Raters were trained using three complete tapes

(nine teaching sessions) representing the teaching sessions

of one S from each of the three treatment groups. In addi-

tion to these, the written model along with discarded ver-

sions of the film-mediated model were also used for train-

ing purposes. The experimenter and team of raters system-

atically rated each of the training scripts and tapes.

On the basis of experience with these, the preliminary

manual was revised, and a final Rater Manual was developed

(Appendix A). Approximately 40 hours were spent in training.

Once an acceptable level of agreement was reached,

the raters independently rated the experimental scripts.

Reliability was maintained throughout the analysis by

scheduling occasional meetings where raters would review

relevant decision rules and training scripts. This was

done to insure that systematic rating biases would not

54

develop. Neither ratings done during training nor in sub-

sequent meetings were used in the statistical analysis of

results.

Rating Procedures

Scripts were typed from audio tapes of each teach-

ing session and were coded according to SI treatment and

number of teaching session. This was done to insure that

the raters had no knowledge of the Ss name, treatment con-

dition, or phase in the treatment condition. In each

script, all questions were subsequently marked and numbered.

Scripts were than subdivided into four groups and distri-

buted to raters. When scripts were completed by one judge

they were exchanged with the three other judges until each

judge had rated every script. Each rater used a standard

form (Appendix E) for recording the occurrence o° Analytic

Questioning behavior. Rater scores were subsequently

transferred to coding sheets and data cards for analysis.

Rater Reliability

The reliability .of scores was established for

judges 1, 2, 3, and 4 over all rated categories. The

55

Analysis of Variance repeated measures model described

by Winer (1962) was used for this purpose. This analysis

provides two reliability coefficients: (1) an estimate

of the reliability of the mean of four ratings, which

represents the expected correlation between mean ratings

for the same people with other random samples of judges,

and (2) an estimate of the reliability of a single mea-

surement, which is approximately equal to the average

intercorrelation between ratings given by pairs of judges.

The reliability of both the mean scores and of

a single measurement for each rated category for each

teaching session appears in Table 5. These coefficients

are extremely high, suggesting that the criterion measures

are sufficiently reliable for testing both treatment ef-

fects and ability-performance relationships.

The reliability coefficients reported are some-

what higher than expected on the basis of the general

literature on rating. The extensive pretraining of raters,

the specificity of the behavior rated, and the derivation

of the ratings from static transcripts rather than from

ongoing behavioral sequences are believed to have contri-

buted to the high reliability coefficients.

TABLE 5

RATER RELIABILITY

TeachingSession 1

TeachingSession 2

TeachingSession 3

Category a b a b a

Total Analytic .95 .88 .99 .96 .99 .97

Hypotheses .95 .87 .96 .88 .96 .88

Definitions .94 .85 .97 .90 .97 .94

Assumptions .98 .93 .99 .97 .99 .97

Facts .97-, .90 .97 .94 .97 .93

Conclusions .90 .75 .95 .87 .96 .95

Total Categories .96 .90 .98 .94 .98 .95

High Quality Level .92 .81 .98 .94 .98 .95

Total Non Analytic .98 .95 .99 .97 .99 .96

a - Figures in this column representmean scores.

coefficients of the reliability of

b - Figures in this column representsingle measurement.

coefficients of the reliability of a

CHAPTER III

RESULTS

The primary objectives of this study were:

1. To assess the relative effects of different

modes of model presentation.

2. To explore the effects of individual differ-

ences on observational learning.

This chapter will describe the statistical tests

of the hypotheses and the results achieved. The presen-

tation of results will treat first, the instructional

treatment main effects; second, the ability measures;

and finally, the aptitude x treatment interactions.

A complete list of all independent and dependent

measures to be analyzed appears in Table 6. The distribu-

tions of scores for the dependent measures permitted evalua-

tion by parametric techniques. Most analyses were computed

using UCLA Bio-Medical Programs. (Dixon, 1964)

57

TABLE 6

INDEPENDENT AND DEPENDENT MEASURES

58

Independent Measures Dependent Measures

Hidden Figures

Hidden Figures

Hidden Figures

Identical Pictures

Identical Pictures

Identical Pictures

Maze Tracing

Maze Tracing

Maze Tracing

Word Arrangement

Word Arrangement

Word Arrangement

- Part I

- Part II

- Total

- Part I

- Part II

- Total

- Part I

- Part II

- Total

- Part I

- Part II

- Total

Advanced Vocabulary - Part I

Advanced Vocabulary - Part II

Advanced Vocabulary - Total

Verbal GRE

Film Memory

Memory for Ideas -

Memory for Ideas -

Memory for Ideas -

Sentence Reproduction-



Part I

Part II

Total

Part I

Part II

Total

Analytic Questions, T1

Categories of Analytic Questions, T1

High Quality Analytic Questions, T1

Nonanalytic Questions, T1









True-False Test

Matching Test

59

Instructional Treatment Main Effects

Classroom Performance Measures

The means of classroom performance measures were

derived for each S by averaging the scores given on each

category by each of the four judges. All analyses were

computed using these mean scores.

The results for the four 'classroom performance

measures previously described were analyzed separately for

each of the three teaching sessions, although in some cases,

the relationship among these measures was sufficiently high

to suggest that they do not represent psychologically dif-

ferent variables. This was done in view of differences in

ability-performance relationships observed and because of

the unavailability of computer programs for multivariate

analysis of variance techniques. Intercorrelations among

dependent measures are reported both for the total sample

(Table 7) and for each experimental group separately (Ap-

pendix F).

The initial test used to determine if there were

significant instructional treatment main effects on the

criterion variables was a 3 x 3 repeated measures analysis

of variance. Since conformity of the data to the homogeneity

TABLE 7

INTERCORRELATIONS AMONG DEPENDENT VARIABLES FOR TOTAL SAMPLE

2 3 4 5 6 7 8 9 10 11 12 13 14

1

2

3

4

5

6

7

8

9

10

11

12

13

14


Categories, T1

High Quality, T1

Nonanalytic, T1


Categories, T2

High Quality, T2

Nonanalytic, T2


Categories, T3

High Quality, T3

Nonanalytic, T3

True-False Test

Matching Test

77 83

68

15

19

18

29

14

20

03

23

15

14

-07

78

29

14

22

01

95

74

-02 29

06 14

-04 23

34 09

-30 74

-36 64

-31 68

-30

22

10

14

-06

70

73

67

-31

83

25 09

10 15

25 11

06 44

73 -2.8

62 -2S

71 -29

033 65

94 -27

81 -27

-31

08

00

-05

06

49

40

45

-21

48

46

48

-24

-07

-07

-16

-16

42

43

42

-21

40

47

42

-29

56

Decimals omitted

61

assumptions underlying the use of this model appeared ques-

tionable (Appendix G), separate one and two-way analyses

of variance were also computed for each measure of the

dependent variable. Since the results were the same for

both methods of analysis, the repeated measures analysis

of variance was selected as the more parsimonious way of

reporting these data. However, the results of the separate

one and two-way analyses of variance are reported in Ap-

pendix H.

In accordance with Winer (1962), when the inter-

action between the two factors in the repeated measures

analysis of variance was significant, tests on simple main

effects were computed in addition to direct tests on main

effects. In such cases, both direct tests on main effects

and tests on simple main effects are reported separately.

In all cases, the Newman-Keuls procedure (Winer, 1962) was

used in comparisons of pairs of treatments following a

significant overall F ratio (Appendix I).

Table 8 presents the means and standard deviations

of the classroom performance measures for each of the three

experimental groups.

Analytic Questions. The results for the repeated

measures analysis of variance of the frequency of Analytic

TABLE 8

MEAM_AND STANDARD DEVIATIONS OF DEPENDENT. VARIABLES

Performance Measure

Treatment GroupFilm - Mediated

ModelingWrittenModeling Control

Mean S.D. Mean S.D. Mean S.D.

Analytic Questions, T1 3.61 3.37 3.62 3.30 3.72 3.10Categories of Analytic Questions, T1 1.43 1.12 1.75 1.01 1.76 1.11High Quality Analytic Questions, T1 1.88 2.02 1.75 1.89 2.18 2.24Nonanalytic Questions, T1 23.26 9.41 24.39 9.33 25.24 10.80Analytic Questions, T2 13.40 6.41 8.16 5.59 3.31 2.50Categories of Analytic Questions, T2 3.73 1.19 2.89 1.39 1.59 1.05High Quality Analytic Questf.ons, To 9.90 5.20 5.17 3.84 1.92 1.70Nonanalytic Question, T2 ` 16.02 9.36 19.12 11.49 25.40 10.57Analytic Questions, T3 13.58 6.33 10.09 7.11 2.93 3.25Categories of Analytic Questions, T3 3.85 1.17 2.98 1.37 1.33 .97High Quality Analytic Questions, T3 10.15 5.52 6.72 4.42 1.68 2.24Nonanalytic Questions, T3 15.50 8.16 17.23 9.15 25.96 12.72Matching Test 11.82 2.36 10.67 3.09 7.97 2.88True-False Test 10.70 2.02 9.77 2.35 7.00 1.90

63

Questions are presented in Table 9. In view of the signif-

icant interaction between treatment groups and teaching

sessions, (F = 29.41, p < .01) tests on simple main effects

were computed (Table 10). These tests disclosed highly

significant treatment effects both for T2 (F = 46.27, p<.01),

and for T3 (F = 50.55; p < .01), while no differences be-

tween groups were found for T1 (F = .01). Comparisons be-

tween pairs of treatment conditions using the Newman-Keuls

procedure showed that for both T2

and T3, the Written and

Film-Mediated Modeling groups generated significantly more

Analytic Questions (p < .01) than did the Control group.

Moreover, Ss in the Film-Mediated Modeling condition pro-

duced significantly more Analytic Questions than did Ss

in the Written Modeling condition on both T2 and T3

(p < .01).

An alternative view of the relative efficacy of

the different instructional treatments in augmenting the

frequency of Analytic Questions is furnished by within-

group analyses of the increase from base rate to subsequent

teaching sessions. The increase was found to be significant

for both the Film-Mediated (F = 30.13; p < .01) and Written

(F = 9.74; p < .0.1) Modeling groups. The Control group did

not display increments in the frequency of Analytic Ques-

tions (F = .14). These modeling effects are illustrated in

Figure 2.

64

TABLE 9

ANALYSIS OF VARIANCE OF ANALYTIC QUESTIONS

Source of Variation df MS

Between Subjects 119

Treatment Groups 2 1422.66 32.10 .01

Subjects Within Grow. 44.32

Within Subjects 240

Teaching Sessions 2 992.67 75.31 .01

Treatments X Sessions 4 387.64 29.41 .01

Sessions x Subjects Within Groups 234 13.18

65

TABLE 10

SUMMARY OF ANALYSIS OF VARIANCE OF ANALYTIC QUESTIONSSIMPLE MAIN EFFECTS

Source of Variation cif MS

Beteen Groups

Treatments at Teaching Session 1 2 .16 .01 NS

Treatments at Teaching Session 2 2 1030.25 46.27 .01


Error Between Groups 351 23.56

Within Groups

Sessions of Film-Mediated Modeling 2 1335.51 30.13 .01

Sessions of Written Modeling 2 442.00 9.74 .01

Sessions of Control 2 6.30 .14 NS

Error Within Groups 234 13.18

.... mot

Amwwwwww====m-,47

14

12

I0

66

MOM

MID

FILM -MEDIATEDMODELING

WRITTEN MODELING

I

TI

CONTROL

T2 T3

FIGURE 2. MEAN FREQUENCY OF ANALYTIC QUES-TIONS ACROSS TEACHING SESSIONS.

67

Categories of Anal tic Questions. Table 1.11' pre-

sents the results for the repeated measures analysis of

variance for the variety of categories of Analytic Questions

used. Since there was a significant interaction (F = 27.65,

p <.01) between treatment groups and teaching sessions,

tests for simple main effects were also performed (Table 12).

These analyses yielded significant treatment effects both

for T2 (F = 34.23; p <.01) and for T3 (F = 48 .21; p <.01).

Differences between treatment groups were not found for

T1 (F = .97). The specific differences contributing to

the treatment effects were therefore investigated by com-

parisons of pairs of treatments. The comparisons showed

that both the Written and Film-Mediated Modeling groups

used significantly more categories of Analytic Questions

both for T2

and T3

(p < .0 ) than did the Control group.

In addition, Ss in the Film-Mediated Modeling condition

used significantly more categories of Analytic Questions

(p < 01) than did Ss in the Written Modeling condition

for both teaching sessions.

Within-group analyses of change for each

group between T1 and T3 showed a significant increase in

the number of categories of Analytic Questions used both

for the Written Modeling group (F = 21.81; p.01) and

68

TABLE 11

ANALYSIS OF VARIANCE OF CATEGORIES OF ANALYTIC QUESTIONS

Source of Variation df MS F p



Subjects Within Groups 117 2.36

Within. Subjects 240



Sessions X Subjects Within Groups 234 .87

TABLE 12

SUMMARY OF ANALYSIS OF VARIANCE OF CATEGORIES OF ANALYTIC QUESTIONS

SIMPLE MAIN EFFECTS

69


Between Groups

Treatments at Teaching Session 1 2 1.34 .97 NS




Within Groups



Sessions of Control 2 1.84 2.11 NS

Error Within Groups 234 .87

70

FILM-MEDIATED MODELIN

WRITTEN MODELING

CONTROL

T3

FIGURE 3. MEAN FREQUENCY OF CATEGORIES OF

ANALYTIC QUESTIONS ACROSS TEACH-

ING SESSIONS.

71

for the Film-Mediated group (F = 86.95; p <. 01), while

the Control group did not show increases in the number

of categories used (F = 2.11).

Figure 3 depicts the mean number of categories

used by each group for each teaching session.

High Quality Analytic Questions. Results from

the repeated measures analysis of variance of the quality

of Analytic Questions appear in Table 13. Testri for simple

main effects were performed (Table 14) in view of the

significant interaction (F = 32.61: p<. 01) between treat-

ment groups and teaching sessions. These tests disclosed

highly significant treatment effects for T2 (F = 51.37;

p<.01) and for T3 (F = 57.80; p<.01), while treatment

effects were not found for T1 (F = .08). Comparisons

b'itween pairs of treatment conditions showed that Ss in

both the Written and Film-Mediated Modeling groups generated

significantly more high quality Analytic Questions than

did Ss in the Control group, both for T2 and for T3 (p <.01).

Similarly, the two modeling procedures proved to be dif-

ferentially effective, with Ss in the Film-Mediated Modeling

group producing significantly more high quality Analytic

Questions than did Ss in the Written Modeling group (p<.01)

for both teaching sessions.

ANALYSIS OF VARIANCE OF HIGH QUALITY ANALYTIC QUESTIONS

Source of Variation

72

F p

Between Subjects

Treatment Groups

Subjects Within Groups

Within Sublects

Teaching Sessions

Treatments X Sessions

Sessions X Subjects Within Groups

119

2 874.83 38.67 .01

117 22.62

240

2 614.88 79.74 .01

4 251.48 32.61 .01

234 7.72

TABLE 14

SUMMARY OF ANALYSIS OF VARIANCE OF HIGH QUALITY ANALYTIC QUESTIONS

SIMPLE MAIN EFFECTS


Between Groups

Treatmentsat Teaching Session 1 2 .90 .08 NS

Treatmentsat Teaching Session 2 2 652.46 51.37. .01

Treatmentsat Teaching Session 3 2 734.11 57.80 .01


Within Groups



Sessions of Control 2 1.33 .32 NS


IO

8

6

4

2

0

FILM-MEDIATED MODELING

WRITTEN MODELING

CONTROL

I I I

74

Ti T2 T3

FIGURE 4. MEAN FREQUENCY OF HIGH QUALITYANALYTIC QUESTIONS ACROSS TEACH-ING SESSIONS.

75

Within-group analyses provided another view of

the differential effectiveness of the instructional treat-

ments in producing high quality Analytic Questions. Ss in

both the Film-Mediated Modeling group (F = 117.39; p < .01)

and Written Modeling group (F = 29.25; p < .01) displayed

significant increases in the number of high quality Analy-

tic Questions while Control group Ss did not display signif-

icant increments (F = .32). A visual examination of these

modeling effects appears in Figure 4.

Nonanalytic Questions. It was anticipated that

treatment effects might manifest themselves by a rise in

the frequency of Analytic Questions in some groups and a

related decrease in the frequency of Nonanalytic Questions.

Table 15 presents the results for the repeated measures

analysis of variance of Nonanalytic Questions. Since a

significant interaction between treatment groups and teach-

ing sessions was disclosed (F = 4.47; p < .05), tests for

simple main effects were also computed (Table 16). These

analyses showed significant treatment effects both for T2

(F = 8.04; p < .01) and for T3 (F = 12.08; p < .01). Sig-

nificant treatment effects were not found for T1 (F = .38).

Differences contributing to the treatment effects were in-

vestigated by comparing pairs of treatments. The comparisons

76

TABLE 15

ANALYSIS OF VARIANCE OF NONANALYTIC QUESTIONS




Subjects Within Groups 117 200.59

Within Subjects 240



Sessions X Subjects Within Groups 234 56.48

77

TABLE 16

SUMMARY OF ANALYSIS OF VARIANCE OF NONANALYTIC QUESTIONSSIMPLE MAIN EFFECTS


Between Groups

Treatment at Teaching Session 1 2 39.73 .38 NS

Treatment at Teaching Session 2 2 921.11 8.04 .01

Treatment at Teaching Session 3 2 1262.89 12.08 .01


Within Groups

Film-Mediated Modeling 2 772.13 13.67 .01

Written Modeling 2 550.23 9.74 .01

Control 2 5.60 .10 NS


78

26

24

22

20

18

16

14TI T2 13

FIGURE 5. MEAN FREQUENCY OF NONANALY-TIC QUESTIONS ACROSS TEACH-ING SESSIONS.

79

showed that for T2 and T3, Control group Ss produced sig-

nificantly more Nonanalytic Questions than did Ss in either

the Written or Film-Mediated Modeling conditions (p < .01).

However, the Written and Film-Mediated Modeling groups did

not differ significantly from one another on either occasion.

Differential treatment effects are also shown

by within-group analyses of changes in performance from

T1 to T3. Is in the Film-Mediated Modeling group (F = 13.67;

p < .01) and the Written Modeling group (F = 9.74; p < .01)

displayed significant decreases in the number of Nonanaly-

tic Questions used, while Control group Ss did not display

such decrements (F = .10). It should also be recalled

here that Ss in the Written and Film-Mediated Modeling

conditions displayed significant increases in the frequency

of Analytic Questions, while Control group Ss did not.

These modeling effects are illustrated in Figure 5.

Written Measures

As previously indicated, performance on two

written measures was obtained in addition to classroom

performance measures. The written measures consisted of

a True-False Test in which Ss were asked to identify examples

80

and nonexamples of major categories of Analytic Question-

ing; and a Matching Test, in which subjects were asked

to match correctly given questions with categories of

Analytic Questions.

A one-way analysis of variance was computed be-

tween groups to determine if there were significant in-

structional treatment main effects for the written mea-

sures. Following significant F ratios, the Newman-Keuls

procedure was used in comparisons of specific pairs of

treatments (Appendix (T). Table d presents the means and

standard deviations of the written measures for each of

the three experimental groups.

True-False Test. The results for analysis be-

tween treatment groups of scores on the True-False Test

are shown in Table 17. A one-way analysis of variance

computed between the three groups disclosed significant

treatment effects (F = 20.76; p < .01).

Ss in the Written and Film-Mediated Modeling

conditions correctly identified significantly more examples

and nonexamples of categories of Analytic Questions than

did Ss in the Control group (p < .01). However, scores

for Ss in the Written and Film-Mediated Modeling condi-

tions did not differ significantly from one another.

TABLE 17

ANALYSIS OF VARIANCE OF SCORES FOR TRUE-FALSE TEST

Source of Variation df MS P

Between Croups 2 162.24 20.76 .01

Within Groups 117 7.81

..111

81

Matching Test Table 18 presents the results

for analysis between treatment groups of scores for the

Matching Test. This analysis showed highly significant

treatment effects (F = 32.94; p< .01).

Comparisons between pairs of treatment conditions

showed that Ss in the Written and Film-Mediated Modeling

conditions correctly matched significantly more items on

the test than did Ss in the Control group (p < .01). More-

over, Ss in the Film-Mediated Modeling conditioti obtained

significantly higher scores (p < <05) than did Ss in the

Written Modeling condition.

82

TABLE 18

ANALYSIS OF VARIANCE OF SCORES FOR MATCHING TEST

Source of Variation df MS F P

Between groups

Within Groups

2 145.38 32.94 .01

117 4.41

Ability Measures

Scores on a series of nine ability measures were

obtained for 120 Ss included in the analysis. Following the

completion of testing, Ss were randomly assigned to the three

experimental groups. Table 19 presents the means and stan-

dard deviations of these ability measures for the total sample

in addition to the composition of the three groups separately.

Frequency distributions of the ability measures

were judged to be approxinately normal. Thus, normality was

assumed for the purposes of further analysis.

Reliability

Reliabilities for all tests used, with the excep-

tions of Film Memory and the Graduate RecordsExamination,

TABLE 19

MEANS AND STANDARD DEVIATIONS OF ABILITY MEASURES

Treatment Group

Film-Mediated WrittenModeling Modeling Contr )1

TotalSample

Mean S.D. Mean S.D. Mean S.D. Mean S.D.

Hidden Figures - Part IHidden Figures - Part IIHidden Figures - TotalIdentical Pictures - Part IIdentical Pictures - Part IIIdentical Pictures - TotalMaze Tracing - Part IMaze Tracing - Part IIMaze Tracing - TotalWord Arrangement - Part IWord Arrangement - Part IIWord Arrangement - TotalVerbal Comprehension - Part IVerbal Comprehension - Part IIVerbal Comprehension - TotalVerbal G.R.E.Film MemoryMemory for Ideas - Part IMemory for Ideas - Part IIMemory for Ideas - TotalSentence Reproduction - Part ISentence Reproduction - Part IISentence Reproduction - Total

6.34 4.00 6.47 3.46 6.10 3.92 6.31

6.14 3157 6.87 3.; 63 6.21 3..59 6.4112.48 6.75 13.35 6..17 12.31 7.01 12.741.26 8.09 41.70 5.50 40.68 6.54 41.2339.83 8.13 41.42 6.56 39.47 6.80 40.2580.71 14.90 82.62 12.05 79.95 12.27 81.1110.10 3.45 11.15 3.18 10.86 4.10 10.7013.46 4.07 14.13 3.77 14.63 4.25 14.0623.56 6.98 25.27 6.46 25.55 7.94 24.7719.82 6.71 20.97 6.62 20.18 7.16 20.3322.99 7.83 24.30 7.83 21.11 7.54 22.5541.78 13.63 45.02 13.80 40.89 13.83 42.5913.68 2.92 14.10 2.58 14.03 2.44 13.9314.34 3.19 14.35 2.51 14.21 3.07 14.3027.98 5.76 28.48 4.71 28.23 5.18 28.20

611.95 97.06 582.75 158.00 582.36 139.71 592.6920.34 3.96 20.80 4.70 19.97 3.12 20.3823.00 6.57 22.90 7.87 22.47 6.04 22.8014.93 5.11 16.05 5.71 13.82 4.35 14.9537.85 10.80 38.45 10.20 36.29 9.06 37,55136.93 17.83 140.85 15.50 137.65 12.49 138.48139.29 26.69 143.07 30.56 138.23 18.27 140.23276.22 41.01 281.89 24.27 275.89 26.17 278.03

3.773.586.616.767.20

13.103.594.037.136.797.78

13.752.652.915.19.

133.273.986.83,5.1410.0315.4625.6734.36

84

were estimated from a product-moment correlation between

scores on two separately timed parallel parts of each test.

The Spearman-Brown formula was used to obtain the reliabil-

ity coefficients for the full tests.

KR20 estimates of realiability for the Graduate

Records Examination were available in the manual provided

by test publishers. Reliability of Film Memory was also

computed using the Kuder-Richardson formula.

Reliabilities for all tests are provided in

Table 20, along with a standard error of measurement com-

puted for each ability measure. Reliabilities of factor

tests from the French Kit were consistently higher than

those of other factor tests used. Reliabilities for the

audio-visual tests were acceptable, although somewhat lower.

Intercorrelations ofAbility Measures

Intercorrelations of ability measures were com-

puted both for the total sample (Table 21) and for each

of the three experimental groups separately (Appendix J).

For the most part, the factor tests display low

correlations with one another. This is in accordance with

85

TABLE 20

RELIABILITIES AND STANDARD ERRORS OF MEASUREMENTFOR ABILITY MEASURES

Test Reliability

Standard Errorof Measurement

Hidden Figures (Cf-1) .76 3.23

Identical Pictures (P-3) .83 5.37

Maze Tracing (Ss-1) .86 2.70

Word Arrangement (Fe-3) .82 5.77

Advanced Vocabulary (V-4) .87 1.86

Memory for Ideas .65 5.91

Sentence Reproduction .69 19.24

Graduate Records Examination(Verbal Score) .92 28.0u

Film Memory .50 2.98...

TABLE 21

INTERCORRELATIONS OF ABILITY MEASURES:

TOTAL SAMPLE

12

34

56

78

910

11

12

13

14

15

16

17

16

19

20

21

22

23

1.

2.3,

4.

5.

6.

7.8.9.

10.

11.

12.

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

Hidden Figures - Part I

Hidden Figures - Part II

Hidden Figures - Total

Identical Pictures - Part I

Identical Pictures - Part II

Identical Pictures - Total

Maze Tracing - Part I

Maze Tracing - Part II

Maze Tracing - Total

Word Arrangement - Part I

Word Arrangement - Part II

Word Arrangement - Total

Verbal Comprehension - Part I

Verbal Comprehension - Part II

Verbal Comprehension - Total

Verbal GRE

Film Memory

Memory For Ideas - Part I

Memory For Ideas - Part II

Memory For Ideas - Total

Sentence Reproduction - Part I

Sentence Reproduction - Part II

Sentence Reproduction - Total

61

9089

34

3337

33 42

4171

34

4142

88 93

3125

32384345

2829

31

434549

75

32

2934

4347

50

9294

2525

28

210815

071612

1014

131820

221019

1668

1720

2120

15

2010

20

1689

93

06 10

09

27 25

2909

1714

1320

18

15

1316

32

3335

2019

211019

1777

12

13

13

31

3134

161919

122019

9394

150712

2724

3030

2328

13161748

5756

-04

1305

-04

06

021609

1314

0610

-02060203

0815

1220

1217

13

04

0924

21 23

070909

08

27

-020702

23182019

12

1614

20

18

11111210

33 48

04

1309

25

16

2117

10 1424

2526

1112

1310

37 90

75

0101

01 1609

11

08

0205

08

03072432

2925

02

28

25

31

0608

0719

131610

05

08

0709

101930

27

30

0131 15

28

53

0205

0419

1215

09

0205

08

0609

2536

3332

02

3422

3384

87

Decimals omitted.

87

the expectations regarding correlations among factor tests

presumed to be independent. However) those factor tests

from the French Kit which may be regarded as measures of

perceptual abilities (Hidden Figures, Identical Pictures,

Maze Tracing) tend to correlate more highly with one another

than they do with other tests. These correlations, while

low, are consistently significant. Measures of verbal

performance display a somewhat weaker tendency to cor-

relate more highly with one another than with primarily

perceptual measures.

Relation of Ability Measuresto Performance

The major purpose of this study, it will be re-

called, was to examine the effects of individual differences

on observational learning. In terms of the model previously

proposed by Melton (1967) for investigating individual dif-

ferences in learning, ability-performance relationships

may be expected to vary as a function of the nature of

the task. It was anticipated that the requirements of

the Written and Film-Mediated Modeling conditions were

sufficiently different to produce different ability-


88

A first step in evaluating these theoretically

expected relationships was to compute for each group sepa-

rately, the correlations between all ability and all per-

formance measures. Both part scores and total scores were

used for this purpose. These correlations appear in Tables

22, 23, and 24. Upon inspection it was apparent that in

several instances the correlations between ability and

performance measures varied substantially across treatment

conditions. The most promising variables (i.e., those

variables which demonstrated the greatest variation among

the groups) were then selected for the investigation of

aptitude X treatment interactions. Only variables which

showed correlations significantly different from zero were

used in this analysis.

It should be noted that in several instances,

there are substantial differences between the correlations

of part scores of a single ability measure with a given

performance measure. The most notable example of this

can be seen in the part scores for Hidden Figures, where

it is obvious that Part I and Part II scores are predict-

ing differentially. Since ultimately we are interested

in predicting learning, in cases such as these, part scores

rather than total scores have been used in subsequent

analysis.

TABLE 22

CORRELATIONS OF ABILITY MEASURES WITH PERFORMANCE MEASURES

FILM-MEDIATED MODELING GROUP

T1

T2

Hidden Figures - Part I 03 08 02 20 -15 -20

Hidden Figures - Part II -13 -14 -11 17 10 -14

Hidden Figures - Total -05 -02 -05 21 -04 -19

Identical Pictures - Part I 22 35 13 14 -13 -11

Identical Pictures - Part II 34 34 32 11 13 -02

Identical Pictures - Total 31 36 24 17 03 -06

Maze Tracing - Part I 00 18 03 07 -10 -09

Maze Tracing - Part II 11 21 25 18 -12 -25

Maze Tracing - Total 06 21 16 14 -12 -19

Word Arrangement - Part I 19 22 12 37 -05 -21

Word Arrangement - Part II 21 21 17 40 29 -09

Word Arrangement - Total 23 26 17 44 14 -17

Advanced Vocabulary - Part I 28 34 11 33 21 00

Advanced Vocabulary - Part II 24 30 07 30 20 03

Advanced Vocabulary - Total 27 34 10 34 20 01

Verbal GRE 36 43 16 15 17 -05

Film Memory -02 01 -01 04 09 01

Memory for Ideas - Part I 27 19 09 09 17 21

Memory for Ideas - Part II 41 32 35 07 17 08

Memory for Ideas - Total 41 36 26 11 18 17

Sentence Reproduction - Part I 30 14 15 21 22 00

Sentence Reproduction - Part II 26 11 13 29 38 19

Sentence Reproduction - Total 30 13 15 28 34 12

Decimals Omitted

r = .26, p < .05

r = .36, p < .01

T34)

CD

E-1

04)Cif

rob

4-1

.0ttO4-1

4-1

fI

00

04-1

Ha0

l

1:C

4-1Cif0

4--1

.0tIO

4-1

04--1

H0Cif

cts

00

-20 05 -25 -49 -20 00 11

00 18 15 -19 06 -04 16

-12 13 -07 -39 -09 -01 15

-09 27 -13 -10 -14 16 03

13 04 08 05 06 02 04

06 08 02 02 00 10 04

-18 21 -13 -14 -21 lg -22

-16 23 -07 -13 -08 28 -13

-18 24 -10 -15 -15 25 -18

-13 40 04 -06 01 56 13

26 39 21 13 21 47 26

08 45 15 03 13 56 21

11 45 12 14 -01 38 23

11 39 05 02 -01 32 33

10 43 09 07 -02 36 31

07 22 -01 -07 -08 27 43

05 -13 39 34 .44 -03 05

11 11 06 -10 03 02 07

17 20 10 12 14 07 10

15 18 07 -02 07 03 11

11 12 -11 -23 -17 16 28

34 27 14 -09 09 14 19

27 23 04 -16 -02 17 24

89

16

30

25

18

30

28

01

00

00

12

30

23

11

28

21

28

-25

23

06

19

-05

21

11

90

TABLE 23


WRITTEN MODELING GROUP

T1

T2

T3Da

.0 0 .00 H H 0 T 4

AlP1-44-4t O2 t O

c 4.) 00 Z at 4.) ' 40Z as ri 0 0 as

G4 c..) = aC 0 =

0)4.) 0

t0 .g..4 HW H

g-4 $.10H9-4tliO CD' 2CZ 4.) 00 Zg as r-1 0

G4 0 = M

.0 W0 E-

0 WW4 E

Al0HZas2 9-I gil4g 4.) g

al AlE E4

Hidden Figures - Part I 28 34 18 14 38 29 36 -05 46 35 45 01 30 -20

Hidden Figures - Part II -08 11 -09 -02 09 01 05 11 -08 -07 -09 02 12 -02

Hidden Figures - Total 11 26 05 07 26 16 23 03 21 15 20 02 24 -14

Identical Pictures - Part I -02 14 -01 10 C8 08 08 -04 13 -06 16 -04 31 -05

Identical Pictures - Part II -02 21 00 18 03 04 03 18 07 -03 08 07 33 03

Identical Pictures - Total 01 20 02 13 03 02 06 11 05 -08 07 06 34 00

Maze Tracing - Part I 05 01 -01 -30 -02 -09 06 02 10 05 11 -20 25 19

Maze Tracing - Part II 05 12 -1- e3 15 06 19 01 20 24 26 -02 55 41

Maze Tracing - Total 05 07 -07 -28 07 01 14 01 17 16 21 -11 45 34

Word Arrangement - Part I 06 22 -03 06 07 12 03 25 18 25 16 19 09 05

Word Arrangement - Part II -02 13 -04 00 -02 04 -02 14 06 20 10 07 02 02

Word Arrangement - Total -01 16 -07 -02 -01 07 -02 21 07 22 10 12 00 04

Advanced Vocabulary - Part I 19 37 13 26 13 -02 18 00 -01 -19 -09 21 08 -09

Advanced Vocabulary - Part II 21 37 11 13 20 08 19 -02 21 -02 10 09 12 00

Advanced Vocabulary - Total 23 37 14 18 14 01 18 -03 08 -16 -01 15 09 -05

Verbal GRE 23 10 15 -01 -11 -06 -07 08 13 -02 01 06 09 -12

Film Memory -03 08 01 12 08 23 04 20 03 14 -02 02 22 32

Memory for Ideas - Part I 14 18 18 15 17 15 18 -19 07 09 07 -21 08 -24

Memory for Ideas - Part II 15 02 12 22 25 25 22 -08 28 22 23 -22 12 07

Memory for Ideas - Total 12 10 16 20 22 21 22 -17 17 19 15 -24 20 -05

Sentence Reproduction - Part I 15 26 19 -07 29 27 31 -13 19 14 20 -28 25 25

Sentence Reproduction - Part II 24 10 09 -30 14 14 27 -25 13 12 23 -34 22 08

Sentence Reproduction - Total 22 21 18 -24 21 20 29 -23 14 10 21 -38 23 16

Decimals emitted.

r= .26, p< .05

r = .36, p< .01

TABLE 24


CONTROL GROUP

T T2

434-1

cd0

Hidden Figures - Part I 17 07 21 01 -04 -14 09 -14 05 03 -04 09 -15 -02

Hidden Figures - Part II 07 01 02 -09 -18 09 -04 00 -02 -02 03 -11 03

Hidden Figures - Total 13 04 11 01 -07 -17 10 -10 03 00 -04 07 -14 03

Identical Pictures - Part I 18 26 19 16 15 09 31 34 18 19 13 29 14 -09

Identical Pictures - Part II 20 22 16 09 09 04 26 22 13 17 17 29 -06 03

Identical Pictures - Total 19 23 18 14 14 06 32 29 17 18 16 31 04 -06

Maze Tracing - Part I 22 28 24 02 30 32 44 02 37 52 33 30 -15 -11

Maze Tracing - Part II 14 21 11 03 12 11 25 17 28 37 28 19 -11 -24

Maze Tracing - Total 18 25 18 02 22 23 36 11 34 47 32 26 -13 -18

Word Arrangement - Part I 05 04 -04 -03 -11 -18 -09 15 09 -01 04 12 07 26

Word Arrangement - Part II 18 -01 17 14 -13 -21 -05 24 21 05 19 29 -01 20

Word Arrangement - Total 15 04 10 07 -11 -22 -07 20 18 04 14 24 03 22

Advanced Vocabulary - Part I 13 -02 11 -30 33 28 20 -23 17 12 09 -05 -25 -05

Advanced Vocabulary - Part II 22 12 09 -29 49 29 38 -19 09 15 -02 06 -05 08

Advanced Vocabulary - Total 19 07 11 -32 45 30 32 -23 13 15 03 01 -15 02

Verbal GRE 20 10 05 -13 37 23 34 -01 14 27 07 03 02 08

Film Memory 23 17 -02 -27 24 22 28 -05 26 39 18 18 18 30

Memory for Ideas - Part I 20 31 21 -04 12 03 09 -02 35 26 33 04 19 07

Memory for Ideas - Part II 04 05 -06 09 -10 -08 -07 07 16 24 17 05 13 05

Memory for Ideas - Total 15 23 11 01 03 -01 03 02 30 29 30 05 19 07

Sentence Reproduction - Part I -05 -12 -02 -26 -11 -22 -07 -25 04 01 -03 -07 -03 16

Sentence Reproduction - Part II -01 -10 -13 -14 19 05 -13 -07 12 06 05 -12 14 -05

Sentence Reproduction - Total -03 -13 -09 -22 07 -07 06 -17 10 05 02 -12 08 04

Decimals omitted.

r = .26, p< .05

r = .36, p< .01

92

One possible explanation for the differential

functioning of the part scores might center on the nature

of the particular tests being considered. These tests

are both unique in nature and high in difficulty level.

Accordingly, the two parts of the tests can be thought

of as two separate trials in a learning task. It

is conceivable that differences in part scores reflect

psychologically different processes.

Aptitude X Treatment Interactions

Aptitude X treatment interactions were evaluated

by comparison of regression slopes for different treat-

ments, using F tests for heterogeneity of regression.

Tests were subsequently performed to determine the signif-

icance of individual regression slopes for each treatment.

Inherent in the basic theoretical considerations

underlying the investigation of aptitude X treatment inter-

actions is educational adaptation to individual differences

in the learner. Instruction can be adapted only if there

are alternative treatments leading to the same terminal

objective, and only if the regression of criterion scores

on aptitude scores have the disordinal pattern illustrated

in Figure 6 in which the regression line relating aptitude

93

to criterion scores under one treatment intersects the

regression line for the alternative treatment (Cronbach,

1965). Consequently, while the results of all regression

Criterion

Treatment A

Treatment B

Aptitude

Figure 6. Disordinal Interaction

analyses computed are reported, only the results of analyses

in which significant disordinal interactions occurred have

been presented in either graphic form or subsequent dis-

cussion.

A preliminary step in evaluating aptitude

treatment interactions was to compute regression equations

for all ability measures, using scores obtained for both

written measures. Of 48 F tests computed to test hetero-

geneity of regression, 6 significant interactions were

found. This step was taken in order to improve procedures

for selecting variables for the analysis of aptitude X

treatment interactions for classroom performance measures.

Of 21 tests for heterogeneity of regression subsequently

94

computed for the classroom performance measures, 7 signifi-

cant disordinal interactions were found. Thus a consider-

able improvement in selection can be observed.

However, it should be noted that the latter fig-

ure is somewhat arbitrarily low. Tests for heterogeneity

of regression were computed for T1 as well as for T2 and T3,

even though it was clear that significant interactions

were to be found only in the later stages. Moreover, when

it appeared that the strength of a relationship between a

total test score and a performance measures was largely

accounted for by the relationship of one of the part scores

to that measure, only the part score was used for regres-

sion analysis, even though in some cases the total score

would also have produced a significant disordinal inter-

action.

Classroom Performance Measures

Analytic Questions. Results of regression analy-

ses of aptitude x treatment interactions, using the fre-

quency of Analytic Questions as the criterion measure, are

summarized in Table 25. These analyses show that scores

on Hidden Figures--Part I produced significant disordinal

TABLE 25

SIMPLE REGRESSION ANALYSES FOR ANALYTIC QUESTIONS

PredictionTeachingSession

Treatment Group

FFilm-Mediated

Modelinga b

WrittenModelinga b

Controla b

Hidden Figures - Part 1 T1

3.44 .03 2.05 .24 2.90 .14 .64

Hidden Figures Part 1 T2

14.90 -.24 4.19 .61** 3.34 .02 4.02*

Hidden Figures - Part 1 T3

16.08 -.3.i)* 3.87 .96** 2.71 .05 8.25**

Film Memory T2

10.60 .14 6.12 .10 -.47 .18 .04

Film Memory T3

.96 .62 9.14 .05 -2.57 .28 1.83

Sentence Reproduction-Part 1 T2

2.66 .08 -6.66 .11 6.22 -.02 1.19

Sentence Reproduction -Fart 1 T3

19.40 -.04 -2.50 .09 1.63 .01 1.37

*p<.05

**p.01

Note: a and b are of the form: 7 box

96

2 4 6 8 10 12 14 16

HIDDEN FIGURES -PART I

FIGURE 7. INTERACTION OF HIDDEN FIGURES - ISCORES WITH INSTRUCTIONAL TREAT-MENTS FOR ANALYTIC QUESTIONS.

interactions with the frequency of Anal tic Questions both

for T2 (F = 4.02; p < .05) and for T3 (F = 8.25; p < .01);

while the interaction for T1 was insignificant (F = .64).

In each case, scores for Hidden Figures were

positively related to performance in the Written Modeling

condition while negatively related to performance in the

Film-Mediated Modeling condition. Thus high ability Ss

learned to use Analytic Questions better from the Written

Modeling treatment than from the Film-Mediated Modeling

treatment, while those scoring low on Hidden Figures profited

more from the Film-Mediated treatment. Hidden Figures

scores did not display consistent relationship to per-

formance in the Control group. It should be noted that

the magnitude of the interactions increased across teaching

sessions. The obtained relationship between scores on

Hidden Figures--Part I and the frequency of Analytic Ques-

tions for the three experimental conditions is illustrated

in Figure 7.

Categories of Analytic Questions Aptitude X

treatment interactions obtained using the variety of cate-

gories of Analytic Questions as dependent variable are

shown in Table 26. The results of these regression analyses

TABLE 26

SIMPLE REGRESSION ANALYSES FOR CATEGORIES OF ANALYTIC QUESTIONS

Prediction

TeachingSession

Treatment Group

Film-MediatedModelinga b

WrittenModelinga b

Controla b

Hidden Figures-Part 1 T1

1.29 .02 1.10 .10 1.64 .02


4.11 -.06 2.14 .12* 1.76 -.04


4.76 -.14** 2.08 .14* 1.30 .01

Film Memory T2

.09 .07 1.49 .07 .09 .07

Film Memory T3

1.05 .12 2.11 .04 -1.05 .12


3.77 .00 -.51 .02 4.06 -.02


5.91 -.02 1.16 .01 1.22 .00

*p<.05

**p<.01

No :e: a and b are of the form: Y = a + bx

F

.97

3.28*

8.88**

.63

.78

2.37

1.54

99

..................................................................................................................

4 6 8 10 12 14 16

HIDDEN FIGURES - PART I

FIGURE 8. INTERACTION OF HIDDEN FIGURES - ISCORES WITH INSTRUCTIONAL TREAT-MENTS FOR CATEGORIES OF ANALYTICQUESTIONS.

100

disclose the interaction of scores on Hidden Figures- -

Part I with instructional treatment conditions both for

T2 (F = 3.28; p < .05) and for T3 (F = 8.88; p < .01).

Hidden Figures scores did not produce significant inter-

action for T1 (F = .97).

Scores for Hidden Figures were positively related

to performance in the Written Modeling condition, while

in the Film-Mediated Modeling condition, a negative re-

lationship was obtained between aptitude and outcome.

Subjects scoring high on Hidden Figures learned to use a

greater variety of Analytic Questions from the Written

Modeling condition; those scoring low benefited more from

the Film-Mediated Modeling condition. Hidden Figures scores

showed little systematic relation to performance in the

Control group. The magnitude of the interactions again

displayed a tendency to increase across performance trials.

These interaction effects are illustrated in Figure 8.

High Quality Analytic Questions The results

of regression analyses obtained with the quality of Analytic

Questions as the criterion measure are presented in Table

27. These results show that scores on both. Hidden Figures--

Part I and Film Memory produced significant disordinal

interactions with the quality of Analytic Questions used.

1

TABLE 27

SIMPLE REGRESSION ANALYSES FOR HIGH .QUALITY ANALYTIC QUESTIONS

Prediction

Treatment Group

Teaching Film-Mediated WrittenSession Modeling Modeling Control F

a b a b a b

Hidden Figures -Part 1

Hidden Figures-Part 1

Hidden Figures- Part 1

Film Memory

Film Memory

Film Memory

Maze Tracing - Part 2

Maze Tracing-Part 2

Maze Tracing-Part 2

T1

1.83 .01 1.43 .10 1.43 .12 .49

T2

11.54 -.26 2.57 .40* 1.63 .04 4.2 0*

T3

11.89 -.27 3.01 .57** 1.88 -.03 5.70*i

T1

2.02 -.01 1.96 .00 2.52 -.02 .01

T2

8.53 .07 4.49 .03 -1.02 .14 .10

T3

-2.38 .62** 7.16 -.02 -.85 .13 4.42*

T1

.21 .12 2.86 -.06 1.33 .06 1.19

T2

12.64 -.20 2.49 .19 .50 .09 1.76

T3

11.61 -.11 2.41 .31 -.48 .15 1.44

*p<.05

**p.01

Note: a and b are of the form: Pi= a + bx

SNOLLS3110 011A1VNV Ainvno HOIH t:104

-143111 1VN011311N.LSNI H11M S3803S I - s3an9u N3001H 40 N01.1.31/831N1 '6 3811914

I leiVd-S3N11914 N300IH 91 01 31 01 8 9 t7 3

1\1111111

ZOT

103

cf) 20zO

w 16a

i= 12

4z4 8

7.144

cD

E

8 12 16 20 24 28 32FILM MEMORY

FIGURE 10. INTERACTION OF FILM MEMORY SCORESWITH INSTRUCTIONAL TREATMENTSFOR HIGH QUALITY ANALYTICQUESTIONS

104

Hidden Figures scores interacted significantly

with the presentation conditions both for T2 (F = 4.20;

p < .05), and for T3 (F = 5.70; p < .01), while the inter-

action for T1 was insignificant (F = .49). These results

were highly similar to those obtained for the frequency

and variety of Analytic Questions, with scores for Hidden

Figures displaying a positive relationship to performance

in the Written Modeling condition and a negative relation-

ship to performance in the Film-Mediated Modeling condition.

Again, Ss with high scores on Hidden Figures learned to

use high quality Analytic Questions better from the Written

Modeling condition, while those scoring low performed better

in the Film-Mediated Modeling condition. Hidden Figures

scores showed little relation to performance in the Control

group. In addition, the interactions tended to increase

as instruction progressed. The relationships obtained

are depicted in Figure 9.

Scores for Film Memory produced a significant

disordinal interaction only for T3 (F = 4.42; p < .05).

Tests for heterogeneity of regression did not disclose

significant interaction either for T1 (F = .01) or for

T2

(F = .10). For T3 Film Memory scores were positively

related to performance in the Film-Mediated Modeling condi-

tion, while unrelated to performance in either the Written

105

Modeling condition or the Control group. These results

show that Ss with high scores on Film Memory learned to

use high quality Analytic Questions better from the Film-

Mediated Modeling treatment, whereas low scoring Ss profited

more from the Written Modeling Treatment. These interaction

effects are shown in Figure 10. In contrast to the inter-

actions previously discussed, this interaction appeared

on T3 with no observable trend in this direction in the

previous teaching sessions.

Written Measures

True-False Test Analyses of the interaction of

ability measures with treatment conditions for scores on

the True-False Test showed that scores for both Film Memory

and Maze Tracing--Part II produced significant disordinal

interactions (Table 28). Although Film Memory scores in-

teracted significantly (F = 4.01; p < .05) with scores

on the True-False Test, the direction of differences was

in contrast to those reported for Film Memory using the

quality of Analytic Questions as the dependent measure.

This time, the regression slope obtained for the Written

Modeling condition was positive, while a negative relation

TABLE 28

SIMPLE REGRESSION ANALYSES FOR TRUE-FALSE TEST

Treatment Grou

FPredictionFilm-MediatedModelinga b

WrittenModelinga

Control

Hidden Figures - Part I 10.06 .11 10.63 .14 7.03 .01 1.03

Hidden Figures - Part II 9.57 .19 9.88 -.02 7.00 .02 .41

Hidden Figures Total 9.62 .09 10.45 -.05 7.00 .00 1.55Identical Pictures - Part I 8.19 .06 10.56 -.02 8.13 -.03 1.10Identical Pictures - Part II 7.81 .07 9.37 .00 6.79 .00 .71

Identical Pictures - Total 7.70 .04 9.71 .00 7.80 .00 1.02

Maze Tracing - Part I 10.75 .00 8.10 .15 11.77 .12 .99

Maze Tracing - Part II 10.74 .00 5.91 .27** 8.58 -.10 4.94**

Maze Tracing - Total 10.74 .00 6.44 .13* 8.14 -.04 3.11*Word Arrangement - Part I 10.17 .03 9.38 .02 5.70 .07 .16

Word Arrangement - Part II 9.13 .07 9.64 .00 6.04 .05 .27

Word Arrangement - Total 9.46 .03 9.46 .00 5.88 .03 .04

Advanced Vocabulary - Part I 9.48 .09 10.99 -.09 7.65 -.04 .59

Advanced Vocabulary - Part II 7.91 .20 9.81 -.00 6.39 .05 .85

Advanced Vocabulary - Total 8.34 .09 10.57 -.03 6.83 .00 .85

Verbal GRE 7.29 .00 10.77 .00 6.48 .00 1.51Film Memory 13.09 -.11 6.32 .17* 3.65 .17* 4.01*

Memory for Ideas - Part I 9.15 .07 11.49 -.08 6.62 .02 2.53

Memory for Ideas - Part II 10.40 .02 9.28 .03 6.78 .02 .00

Memory for Ideas - Total 9.38 .04 10.17 -.01 6.53 .02 .53

Sentence Reproduction-Part I 11.67 -.00 4.56 .04 3.73 .02 1.25

Sentence Reproduction-Part II 8.51 .02 8.81 .00 7.83 .00 .47

Sentence Reproduction-Total 9.20 .00 6.31 .01 6.23 .00 .19

*p<.05**p<.01

Note: a and b are of the form: Ai = a + bx

ILI

Film - MediatedModeling

WrittenModeling

Control

10 15 20 25FILM MEMORY

30 35

FIGURE II. INTERACTION OF FILM MEMORYSCORES WITH INSTRUCTIONALTREATMENTS FOR TRUE - FALSETEST SCORES

14

12

WrittenModeling

Film - Mediated1.. 10 Modelingcowi-w 8Cl)J4U.

1 6Wocc

4

Control

4 8 12 14 20MAZE TRACING - PART II

108

24

FIGURE 12. INTERACTION OF MAZE TRACING -a SCORES WITH INSTRUCTIONALTREATMENTS FOR TRUE -FALSETEST SCORES

109

was obtained between aptitude and outcome in the Film-

Mediated Modeling condition. Thus, high ability Ss learned

to identify categories of Analytic Questions better from

the Written Modeling condition; those scoring low on Film

Memory performed better in the Film-Mediated Modeling con-

dition.

Sao es for Maze Tracing were positively related

to performance in the Written Modeling condition, while

displaying no :elationship to performance in the Film-

Mediated Modeling condition. Subjects scoring high on

Maze Tracing learned to identify categories of Analytic

Questions better from the Written Modeling condition, while

low scoring Ss benefited more from the Film-Mediated Model-

ing condition. The interaction effects obtained are de-

picted in Figures 11 and 12.

Matching Test The evaluation of aptitude

treatment interactions for scores on the Matching Test

showed that both Part II and Total scores for Maze Tracing

displayed significant disordinal interactions (Table 29).

While both scores for Maze Tracing produced significant

interactions, it is clear that the interaction for the

Total score (F = 5.64; p < .01) is largely accounted for

TABLE 29

SIMPLE REGRESSION2tANALYSES FOR MATCHING TEST

Predictcx

Treatment Group

FFilm-Mediated

Modeling

ab

Written

Modeling

ab

Control

a


11.50

.07

8.84

.26*

8.78

-.12

3.33*


11.25

.11

10.49

.00

8.62

-.09

.50


11 26

.05

9.30

.09

8.81

-.06

1.72


11.52

.00

3.47

.17

5.51

.06

1.38


11.15

.02

4.95

.13

9.08

-.03

1.53


11.24

.00

3.89

.08

7.26

.00

1.27


13.44

-.15

7.69

.25

9.21

-.11

2.46

Yaze Tracing - Part II

12.95

-.08

3.77

.47**

9.09

-.07

7.44**


13.40

-.06

4.74

.22**

9.29

-.05

5.64**


10.91

.05

9.87

.03

7.43

.03

.12


9.96

.09

10.70

.00

8.11

.00

1.26

Word Arrangment - Total

10.19

.04

10.51

.00

7.77

.00

.78


9.20

.20

9.23

.09

12.46

-.32

2.43


8.44

.24

7.56

.21

8.76

-.05

1.20


8.28

.13

8.44

.07

10.51

-.09

1.83

Verbal GEE

5.52

.01

9.31

.00

7.79

.00

1.08

Film Memory

11.37

.03

7.44

.15

4.67

.17

.36

Memory for Ideas - Part I

11.36

.02

9.69

.04

6.01

.09

.26

Memory for Ideas - Part II

11.17

.05

9.34

.07

6.83

.09

.03

Memory for Ideas - Total

11.03

.02

8.11

.06

5.74

.06

.28

Sentence Reproduction - Part

I6.63

.04

2.99

.05

9.11

.00

.93


II

9.57

.02

7.03

.02

4.87

.02

.05


7.91

.01

4.29

.02

5.47

.00

.19

*p<.05

**p<.01

Note:

a and b are of the form:

7= a + bx

H 0

111

WrittenModeling

Film MediatedModeling

I0

co

I.- 8CD

6

2

Control

4 8 12 16 20MAZE TRACING - PART /I

24

FIGURE 13. INTERACTION OF MAZE TRACING -ItSCORES WITH INSTRUCTIONALTREATMENTS FOR MATCHING TESTSCORES

112

by the strength of the interaction for Part II scores

(F = 7.44; p < .01). Consequently, since the results are

substantially the same for both scores, only the inter-

action for Part II scores are presented for further dis-

cussion and graphic illustration.

Similar to the results for the True-False Test,

a positive relationship was obtained between aptitude and

criterion in the Written Modeling condition, while scores

for Maze Tracing were unrelated to performance in the Film-

Mediated Modeling condition and the Control group. Again,

Ss with high scores on Maze Tracing learned to identify

specific instances of Analytic Questions better from the

Written Modeling treatment, whereas those scoring low prof-

ited more from the Film-Mediated Modeling treatment. This

relationship is illustrated in Figure 13.

CHAPTER IV

DISCUSSION OF RESULTS AND

IMPLICATIONS FOR FUTURE RESEARCH

Summary of Data: Hypotheses Tests

This study sought to examine the effects of verbal

and perceptual dimensions of individual differences in re-

lation to the efficacy of two different kinds of modeling

proc(Idures in the acquisition of a teaching skill. The

basic hypotheses tested were:

1. Both Film-Mediated and Written Modeling conditions

will produce significantly greater changes in the

response strength of desired behaviors than will

the Control condition.

2. For subjects receiving the Film-Mediated Modeling

condition, criterion scores should show stronger

relation to perceptual abilities than for subjects

receiving the Written Modeling condition.

3. For subjects receiving the Written Modeling con-

dition, criterion scores should show stronger re-

lation to verbal abilities than for subjects

receiving the Film-Mediated Modeling condition.

113

114

The general premises from which these hypotheses

were derived were:

1. The rate and level of learning of a given teaching

strategy varies as a function of model presenta-

tion.

2. The effectiveness of instructionsamethod.s varies

from subject to subject, with such differences

being correlated with trainee aptitudes.

Specific predictions, though tentative, were based

on theoretical considerations which suggested that the re-

quirements of the Written and Film-Mediated Modeling con-

ditions were sufficiently different to produce different

ability-performance relationships.

Instructional Treatment Main Effects

Support for Hypothesis 1 is dependent upon sig-

nificant between-group differences in measures of the de-

pendent variable. The appropriate statistical tests of

this hypothesis are therefore the comparisons of pairs of

treatments following an overall significant F ratio.

Data in Tables 9 through 18 strongly support

this hypothesis. Following initial written instructions

115

common to all groups, both Written and Film-Mediated Model-

ing conditions led to significantly higher frequency, va-

riety and quality of Analytic Questioning than did the

Control group, and thus to a significantly lower frequency

of Non analytic Questions also. Similarly, wiTh respect

to the written measures, Ss in the Written and Film-Mediated

Modeling conditions matched and identified significantly

more correct items than did Control group Ss. Significant

differences existed between the two modeling conditions and

the Control group for all measures of the dependent variable.

It should again be noted that while statistical

consideration has been given to several measures of the de-

pendent variable, the strength of association among these

measures (Table 7 ) in some cases is sufficiently high to

suggest that these measures do not represent psychologically

different variables. However, these variables were ana-

lyzed separately in view of differences which existed in

ability-performance relationships on some occasions, and

because of the unavailability of multivariate analysis of

variance techniques.

Additional support for Hypothesis 1 is provided

by within group analyses of the changes in Analytic Ques-

tioning behavior from base rate to subsequent teaching

sessions. This information is provided in Tables 9 through

18. Both Film-Mediated and Written Modeling conditions

produced significant increases in frequency, variety and

quality of Analytic Questioning., while the Control group

did not display such increases. Moreover, both Written and

Film-Mediated Modeling conditions produced significant de-

creases in the frequency cf Nonanal tic Questioning, while

the Control group did not display such decrements.

The nature of the dependent variable appears

to have been an important factor in the results obtained.

It has been observed that modeling procedures are most

effective in transmitting response patterns which are

new or weakly established (Bandura 1965, 1966). Frequently,

viable teaching skills and strategies, because of their

demonstrable effectiveness in the classroom, can be ex-

pected to be occurring in some strength prior to treat-

ment. This raises the question of the utility of modeling

procedures for training purposes in such instances. How-

ever, the occurrence of Analytic Questioning prior to treat-

ment was found to be surprisingly infrequent. Even follow-

ing exposure to set induction materials describing the na-

ture af analysis, subsequent questioning was most frequently

evaluative rather than analytic. Consequently, the low initial'

117

operant strength of Analytic questioning appears to have

provided optimal conditions 2or demonstrating the effective-

ness of the modeling procedures.

A most striking finding was the consistent su-

periority of the Film-Mediated Modeling condition across

all measures of the dependent variable. No specific pre-

dictions had been formulated concerning differences exist-

ing between 'he Written and Film-Mediated Modeling condi-

tions. While both of these methods have been effective as

training procedures, differences between the two treatments

have not consistently been found (Bandura and Mischel,

1965, McDonald and Allen, 1967, McDonald, 1968). However,

systematic feedback and reinforcement have often been

included in these studies. While the inclusion of these'

variables has generally contributed to strengthening both

modeling treatments, it has perhaps obscured the specific

contributions of the observational treatment variable.

This interpretation of the lack of consistency in

these results would appear to conform to Bandura's argument

that contiguity accounts for the acquisition of matching

responses, whereas reinforcement influences the performance

of imitatively learned responses (Bandura, 1965). Some

evidence on this point has been provided in an experiment

118

in which the introduction of positive incentives com-

pletely wiped out previously observed performance differ-

ences (Bandura, 1965). In contrast, the exclusion of

feedback and reinforcement variables in the present study

permitted investigation of the specific effects of an ob-

servational experience on the acquisition of a teaching

skill.

The clear cut differences between the Written

and Film-Mediated Modeling conditions across all measures

of the dependent variable supports Carroll's (1959) conten-

tion that:

Psychologists have too often confused the spoken andthe written word, or at least they have assumed toofreely that spoken and written words are equivalentstimuli (p. 110).

A typed transcript differs from a filmed portrayal in sev-

eral important respects. A detailed discussion of these

differences was given in Chapter I. In addition to pro-

viding a behavioral conception of both the teaching skill

and lesson pace, the film-mediated portrayal standardizes

intonation and stress patterns which subjects may implicitly

impose. Differences of this nature may prove to be of

considerable importance. Thus, an actual performance is

119

likely to provide substantially more relevant cues with

greater clarity than can be conveyed by a verbal descrip-

tion. From the average data alone then, training under

Film-Mediated Modeling conditions appears to have been

generally more efficient.

Aptitude X Treatment Interactions

Hypotheses 2 and 3 imply disordinal interaction

between aptitude and treatment--more specifically, that

the regression line relating aptitude to criterion scores

under one treatment intersects the regression line for the

alternative treatment. The statistical tests of these hy-

potheses are F tests for heterogeneity of regression.

The evidence with respect to Hypotheses 2 and

3 is mixed. Although aptitude x treatment interactions

were obtained, the direction of the interactions did not

consistently correspond to predictions. Analyses of apti-

tude X treatment interactions indicated that scores on Film

Memory interacted significantly with the quality of Ana-

lytic Questions. Scores for Film Memory were positively

related to performance in the Film-Mediated Modeling con-

dition, while unrelated to performance in the Written

120

Modeling condition. Subjects scoring high on the Film

Memory Test learned to use high quality Analytic Questions

better from the Film-Mediated Modeling treatment, while

those scoring low profited more from the Written Modeling

treatment. These results are consistent with Hypothesis 2.

In contrast to these findings however, scores

on Hidden Figures produced significant disordinal inter-

actions with the frequency, variety and quality of Analytic

Questions. While scores on Hidden Figures were positively

related to performance in the Written Modeling condition,

they were negatively related to performance in the Film-

Mediated Modeling condition. Thus high ability Ss learned

to use greater frequency, variety and quality of Analytic

Questions from the Written Modeling condition than from

the Film-Mediated Modeling condition; those scoring low

on Hidden Figures learned better from the Film-Mediated

Modeling condition. Similarly, with respect to performance

on the written measures, scores on Maze Tracing and Film

Memory interacted with presentation conditions. Using

Film Memory scores, the regression slope obtained for

the Written Modeling condition was positive while a nega-

tive relation was obtained between aptitude and outcome

in the Film-Mediated Modeling condition. Scores for Maze

121

Tracing were positively related to performance in the Writ-

ten Modeling condition, while unrelated to performance in

the Film-Mediated Modeling condition. These results show

that Ss with high scores on Film Memory and Maze Tracing

performed better on the written measures in the Written

Modeling treatment, whereas low ability Ss benefited more

from the Film'-Mediated Modeling treatment. These results

are in sharp contrast to those predicted. Consequently,

Hypotheses 2 and 3 are unsupported by these data. While

aptitude x treatment interactions were obtained, the direc-

tion of the interactions was not consistent and at times

opposite to that hypothesized.

Whenever differences result from a study in con-

tradiction to the hypotheses stated, it is incumbent upon

the experimenter to offer post hoc explanation for such dif-

ferences, although it is not possible, of course, to verify

such speculatory explanation within the scope of this study.

The initial hypotheses were based upon an analysis

of learning tasks and processes corresponding to a model

proposed by Melton (1967) for the investigation of learning

and individual differences. An inspection of the indi-

vidual differences believed to enter at each point in the

learning process can be seen in Figure 14. The organization

rm (sm )

Si1 1(s ) >RaR

Re

* Hidden Figures* Maze Tracing

Identical PicturesVerbal ComprehensionVerbal GRE

122

* Film MemoryMemory for IdeasSentence ReproductionExpressional Fluency

Figure 14. Organization of AbilityVariables in a Multi-Process Model of Learning

of ability measures within the model has been somewhat

arbitrary. Measures such as the GRE Verbal Aptitude scores,

with both comprehension and reasoning components, are

sufficiently diverse to suggest that there is reason to

consider alternative placement. Other measures may also

be characterized by similar overlap. However, it should

be emphasized that this model has been used as a heuristic

device in attempting to select and organize task and ability

variables. It is not intended that the present organization

be accepted as definitive. Further research will be re-

quired to clarify the use of this model.

Of nine ability variables studied, three were

found to interact with presentation conditions. If the

direction of the interactions is examined with respect to

the point in the learning process at which they are believed

t23

to enter, certain trends are suggested. Accordingly, it

is suggested that such trends may conceivably be interpreted

within the framework of the model.

It is becoming increasingly apparent that hypoth-

eses concerning the direction of aptitude g treatment

interactions cannot be based exclusively upon superficial

content similarities between aptitude tests and learning

tasks. Visual or verbal modes of instructional presen-

tation may or may not be related to corresponding scores

on perceptual or verbal aptitude tests.

While it would appear thAt a perceptual mode of

presentation would constitute a demand on perceptual en-

coding systems, the audio-visual presentation could also

conceivably serve a compensatory function through the pro-

vision of perceptual information that might otherwise be

demanded of the subject. Similarly, the lack of audio-

visual content in instructional presentation may require

subjects to generate their own perceptual detail, thus

constituting a demand on the perceptual encoding system.

The stimulus differentiation component of the

model is specifically concerned with the encoding process.

Of the five ability variables studied here, two were found

to interact with presentation conditions. An inspection of

124

the direction of these interactions can be interpreted to

suggest that those abilities involved in the stimulus dif-

ferentiation component may generally serve the compensa-

tory function described above. The purely verbal presen-

tation of a behavioral sequence may require subjects to

generate a behavioral conception of that sequence, thus

creating a perceptual demand. Conversely, an audio-visual

presentation, rich in perceptual detail may require an ab-

straction of a verbal conception, constituting a verbal

requirement.

This compensatory function may be illustrated

in the case of those interactions involving Hidden Figures.

Performance in this test appears to require subjects to

visualize the configurations they are looking for within

the context of a given perceptual field. Knowledge of

the exact configurations which have to be kept in mind

is specially emphasized. Within the Film-Mediated Modeling

condition perceptual schema are provided. Ss low in the

ability to generate such information for themselves are

compensated by this treatment and thereby improve their

performance. However, for Ss who are able to generate such

schema for themselves, the presentation of such information

at a predetermined pace may not only fail to be facilitative.

125

but may inhibit or interfere with the generation of per-

ceptual information or produce frustration and boredom.

An interesting observation was that learners with high

scores on Hidden Figures were actually hindered in the

Film-Mediated Modeling condition, as demonstrated by the

negative relationship between aptitude and performance.

The presence of such negative ability-performance relation-

ships may indicate that reliance on a strong ability may

lead a learner to choose a strategy which unwisely attempts

to exploit that ability in a situation where information

is available which would make a different strategy more

effective (Bunderson, 1968).

In the Written Modeling condition, where per-

ceptual schema are not provided, Ss high in the ability

to generate such information appear to do well, possibly

because following their own pace, they can freely generate

the needed perceptual information. In contrast, Ss low

in ability to generate perceptual information for them-

selves are not provided with this information in the Written

Modeling condition, and subsequently perform less well.

Because of the ',connection of Hidden Figures

with both generalized intellectual ability (Botzum, 1951;

Pemberton, 1952) and with a series of tests of a trait

126

called "field independence" (Witkin et al., 1952), there

is reason to consider possible alternative interpretations

of these results. However, the alternative interpreta-

tions are not necessarily inconsistent with the present

interpretation. Witkin et al. (1962), in a review of the

relevant literature concluded:

These studies provide impressive support for theview that flexibility of closure and field-independencemay be different names for the same dimension. (p. 52)

These findings support Thurstone's (1944) contention that

flexibility of closure may be found to represent parameters

which transcend the immediate perceptual content in terms

of which it had tentatively been identified.

Interactions involving Maze Tracing scores are

believed to be more closely related to the pacing factor.

Maze Tracing requires speed in scanning a complicated

spatial field. In this case, the pace of the film-mediated

model may have served as an equalizer in which the pre-

determined pace of the model imposed a ceiling on the

rate at which Ss high in this ability could scan and process

information, while being sufficiently slow to allow those

Ss low in this ability to profit thereby. Consequently,

the correlation between Maze Tracing and performance

in the Film-Mediated Modeling condition was atten-

uated.

127

In addition, the written model can be charac-

terized as a spatial configuration as well as a verbal

configuration. Teacher discourse was identified and sepa-

rated from student discourse. Questions were set off

by question marks and were generally located in certain

positions within a paragraph of teacher discourse. Cue

discrimination was provided in upper case letters marked

by parentheses. Ss in the Written Modeling condition

could determine their own pace without limitations im-

posed on the speed at which this array of information

could be scanned and reviewed. Accordingly, Ss high in

spatial scanning ability could quickly review this writ-

ten material and locate points at which critical infor-

mation (i.e., cue discrimination and specific examples

of Analytic Questions) was given and thereby improve their

level of performance. Low ability Ss might be less able

to take advantage of these features.

While these two ability variables are the only

variables in the stimulus differentiation component of the

model which interacted significantly with instructional

presentation conditions, patterns of correlations with re-

spect to other ability variables in the stimulus differ-

entiation component are not inconsistent with this

interpretation. Although verbal abilities did not inter-

act significantly with the presentation conditions, the

possibility exists that the cue discrimination, in which

appropriate verbal labels were provided for critical be-

havioral sequences served to attenuate the relationship

between verbal abilities and performance.

Findings such as these on the relationship of

test to task are not unique in the literature on aptitude

X treatment interactions (Blaine and Dunham, 1968). Re-

sults suggesting a compensatory function of certain task

elements have occurred sufficiently often to warrant con-

sidering these data relevant to speculations on the rela-

tionship of a given aptitude to a particular phase in a

learning task.

In contrast to these findings, are the results

obtained regarding the response integration component of

the model. Of the four abilities studied here, two apti-

tude X treatment interactions were disclosed. Both aptitude

x treatment interactions involved Film Memory in which the

129

findings were in direct contradiction to one another. It

will be recalled that the quality of Analytic Questions used

was positively related to FilmIMemory in the Film-Mediated

Modeling condition and unrelated to performance in the

Written Modeling Condition. Conversely, using scores on

the T-F test as the criterion measure, Film Memory was

positively related to performance in the Written Modeling

condition and negatively related to performance in the Film-

Mediated Modeling condition. Although the criteria are

clearly different, such apparently contradictory results

provide an unsatisfactory basis for conclusions regarding

the relationship between the response integration component

of the model and the expected direction of aptitude x treat-

ment interactions. However, a supplemental study of another

set of dependent measures was undertaken. While these

measures were not a part of the main investigation, and thus

will not be reported in detail, the results obtained are rele-

vant to clarification of ambiguities in the data reported.

The supplemental set of dependent measures con-

sisted of a recall test in which Ss were asked to list in

writing as many Analytic Questions as possible, based on

their lesson material, within a 10 minute period. This

was collected along with the other two written measures

44

130

reported) and was rated in a manner identical to that of

the classroom performance measures. Analyses of aptitude

X treatment interactions for these data disclosed that

scores on Film Memory interacted significantly with the

frequency and quality of Analytic Questions. In both cases,

scores on Film Memory were positively related to performance

in the Film-Mediated Modeling condition and unrelated to

performance in the Written Modeling condition. Thus Ss

scoring high on Film Memory asked higher frequency and

quality of Analytic Questions in the Film-Mediated Modeling

treatment) while low scoring Ss performed better in the

Written Modeling treatment. Similarly) scores on Sentence

Reproduction-Part II interacted significantly with the

variety of categories of Analytic Questions used. While

the regression slope obtained for the Film-Mediated Modeling

condition was positive) scores on Sentence Reproduction were

unrelated to performance in the Written Modeling condition.

Subjects with high scores on Sentence Reproduction learned

to ask a greater variety of Analytic Questions from the

Film-Mediated modeling condition; those scoring low bene-

fited more from the Written Modeling condition. These

findings suggest that in contrast to aptitude X treatment

interactions involved in the stimulus differentiation

131

component, aptitude X treatment interactions involving

the response integration and memory components of the

model may generally be consistent with earlier predictions- -

more specifically, the relationship of test to task appears

direct rather than compensatory, with scores on tests of

audio-visual memory related to performance in an audio

visual mode of presentation, while unrelated to performance

in a written mode of presentation.

It is recognized that this interpretation cannot

accommodate the relationship between Film Memory scores

and performance on the T-F test. At this time the experi-

menter is unable to provide a satisfactory explanation

for these results on the basis of present data. While

it is possible that this finding may be due to chance, an

alternative explanation might center on the nature of the

task itself. The discriminations required in the T-F

test are much coarser than those required for other de-

pendent measures. Conversely, the discriminations required

on the Film Memory test are quite fine. An examination

of the pattern of correlations suggests that the relation-

ship of ability measures to performance on the T-F test

is somewhat different from those obtained with other mea-

sures of the dependent variable. It has been suggested

(Fitts, 1965) that the perception and encoding of stimuli

132

depend in part upon the required response mode. However,

the precise effects of such differences in the nature of

a task remain a suitable topic for future investigation.

Information concerning the associational-

mediational component of the model is limited. Within

the context of this study there were no available measures

of ability to generate associative context, analogies,

or verbal-behavioral transformation. However, an experi-

mental measure was developed for use in collecting sup-

plemental data which attempted to assess the ability to

abstract a verbal conception from a behavioral representa-

tion. While scores on this test did not interact signif-

icantly with performance on any of the measures reported in

the main investigation, these scores did interact signif-

icantly with the frequency of Analytic Questions on the

recall measure, disclosing a positive relationship to

performance in the Film-Mediated Modeling condition and

a negative relationship to performance in the Written

Modeling condition. These results correspond to those

obtained for the response integration component of the

model. It should be strongly emphasized, however, that

the limited tryout of both content and procedures in ad-

dition to the lack of reliability and validity data for

this test prohibit the use of this information as anything

other than heuristic.

133

In summary, it has been suggested that the model

proposed by Melton (1967) for the investigation of learn-

ing and individual differences might be used as a plausible

basis for predicting the relationships among learner charac-

teristics and learning outcomes under different instruc-

tional procedures, in the manner described above. While

the evidence on this matter is not overwhelming, and may

be variously interpreted, the model would appear to be

useful, at least in a heuristic sense.

An additional finding of considerable interest

was the tendency for aptitude X treatment interactions

to increase in magnitude across performance trials. This

occurred for all interactions involving classroom per-

formance measures. These findings are in contrast to

those reported by Woodrow (1938), Fleishman and his as-

sociates (1965) and others (Gagne and Paradise, 1961) in

which the contribution of cognitive factors in a learning

task tended to decrease with practice. This has generally

been true both for intellectual and psychomotor tasks.

There are several reasons for such differences.

First of all, the type of task that has generally been

used has been of relatively short and simple content.

This may serve to impose a ceiling on performance which

134

may be reflected in the decreasing contribution of cogni-

tive factors across learning trials. Conversely, the

present study has investigated ability-performance re-

lationships across trials on an intellectually complex

task. In addition, the conditions of performance in the

present study impose a ceiling only in terms of time rather

than a fixed format with a limited number of problems,

frames, or tasks.

Moreover, this line of investigation has general-

ally been concerned 1.7.1.01 task variables rather than in-

structional variables. Accordingly, ability-performance

relationships have been assessed under conditions of prac-

tice rather than instruction. In contrast, the present

study has investigated ability-performance relationships

under different instructional conditions.

Thus, the combination of the type of task and

instructional conditions may have served to increase the

magnitude of aptitude-treatment interactions across trials.

It should be noted, of course, that the number of trials

was limited. Continuation through further teaching sessions

may eventually have resulted in decreasing ability-


135

Implications for Future Research

This study represents an initial attempt to

assess the effects of trainee aptitudes on observational

learning in the acquisition of a teaching strategy. The

results obtained suggest that additional research evaluating

differentiation of instruction in teacher training may

be profitable. The surface has barely been scratched in

this area. A variety of relevant research questions are

suggested by the results of this investigation.

With respect to the use of modeling proceaures

in teacher training, several important issues have yet

to be systematically explored. As Cronbach (1965) has

suggested, any meaningful comparison of treatments re-

quires that each treatment be refined into a good repre-

sentative of its kind. Accordingly, we need to look more

closely at the precise nature of effective modeling con-

ditions in order to determine optimal sequencing of in-

struction and number of cues and exemplars provided, as

well as at the effects of varying amounts and types of

set induction and practice. Moreover, the question of

retention and transfer effects of this training has not

yet been examined. A treatment has multiple effects, and

the effectiveness of instructional methods may differ

136

for immediate mastery, retention and transfer. Although

positive preliminary information has been provided on

this point (Bandura and Mischel, 1965) additional research

is required.

There appears to be an abundance of teacher

behaviors which could ostensibly be used as dependent

variables in studies of this type. Anything that a teacher

does or says in an effort to promote student learning

is a potential skill to be learned by prospective teachers.

While the modeling procedures utilized have been successful

in transmitting specific skills such as Analytic Question-

ing, the general validity of these skills for producing

changes in pupil behavior has yet to be demonstrated.

Although many of these skills possess face validity, at-

tention should be given to the relevance of such teacher

variables in terms of their effect on student learning.

With respect to the effects of trainee aptitude

on observational learning, several research questions are

suggested. The results reported in this investigation

indicate that differences in instructional methods facili-

tated learning in some subjects while inhibiting learning

of others, with such differences being related to trainee

aptitudes. It has been suggested that the model proposed

137

by Melton (1967) for investigating individual differences

in learning might be used as a plausible basis for the

prediction of ability-performance relationships. Con-

siderable emphasis has been placed upon the simultaneous

consideration of task and process variables. Clearly,

additional research is required to evaluate the utility

of this model. The extent to which response integration,

stimulus differentiation, coding and a variety of media-

tional processes may be involved in new learning can be

experimentally manipulated. In view of the limited in-

formation obtained in this study with respect to individual

differences in mediational processes, particular attention

should be given in the future to this component of the

model.

Related to the major question of the utility

of the model are several questions concerning the nature

of the specific task utilized in this investigation. While

the task was relatively complex, the difficulty level was

to some extent attentuated through the provision of set

induction and cue discrimintation. Although these features

are believed to have generally strengthened both modeling

conditions under consideration, the question of the extent

to which the provision of this additional information

138

altered ability-performance relationships that would other-

wise be a function of the exclusive effects of an observa-

tional sequence remains unanswered.

In addition, the results of this study have

raised the question of the effect of response mode on

ability-performance relationships. These relationships

differed substantially from classroom performance to writ-

ten measures. It has been suggested (Fitts, 1967) that

the perception and encoding of stimuli depends at least

partially on the responses to be made to the stimulus

information. Accordingly, it is recommended that further

experimental tests of the model include variations in the

response mode used by subjects.

Moreover, the question of ability-performance

relationships involved in retention and transfer measures

on this particular task has yet to be explored. A legiti-

mate question is the extent to which the battery of tests

used samples abilities related to retention and transfer

as well as immediate mastery.

Finally, further investigation of changes in

the contribution of cognitive abilities over practice

would appear to be a potentially fruitful line of investi-

gation. Results such as those obtained in this study

139

indicate that this question cannot be answered simply.

The effects of variables such as the nature of the task,

conditions of instruction or practice, length of the train-

ing period and composition of the subject group appear

worthy of future investigation.

Implications for Educational Practice

The results of this investigation support the

initial premises that the rate and level of learning of

a specific teaching strategy varies as a function of model

presentation; and that the effectiveness of instructional

methods varies from subject to subject, with such dif-

ferences being related to trainee aptitudes.

Although it has not been demonstrated that the

training procedures utilized in this investigation repre-

sent the most effective way to train teachers, these find-

ings provide evidence that through observation, trainees

can acquire principles exemplified in a model's behavior

and use them for generating novel combinations of teaching

behavior. The results of these training procedures were

sufficiently strong and consistent to suggest that they

can be used with reasonable confidence in their effec-

tiveness. Accordingly, this research supports the general

140

recommendation that the use of written and film-mediated

modeling procedures is a highly effective means of modifying

teaching behavior in training contexts analagous to those

described in this experiment. Moreover, while questions

concerning interactions between specific teaching behaviors

and instructional conditions have yet to be resolved,

there is further evidence, from the average data alone,

to suggest that film-mediated modeling procedures possess

cueing properties which tend to recommend their use over

written modeling procedures, at least for establishing

Analytic Questioning. behavior.

This research does not lead to detailed sugges-

tions for specific modification and developments regarding

the individualization of teacher training programs. Ex-

periments such as these have only begun to explore the

wide range of problems concerned with finding effective

teaching techniques for students with different character-

istics. Much research remains to be done before such

recommendations can be made with confidence. However,

there are both cost and efficiency implications for teacher

training in these results. Given replication of these

findings, assignment of trainees to alternating treatment

is appropriate for maximizing learning. Moreover, the

141

cost of training is reduced considerably as more teachers

can be-assigned to written rather than film-mediated treat-

ments.

The results obtained in this study appear suf-

ficiently encouraging to suggest the potential value of

further research evaluating differentiation of instruction

in teacher training. Results such as these, if replicated,

may eventually provide a basis for the individualization

of teacher training programs.

Limitations and Further Statistical Considerations

While the initial purposes of this investigation

have been satisfied, the results obtained generate questions

concerning the most appropriate kinds of statistical anal-

yses for research of this type. The analyses used have

not been exhaustive. More complete statistical elabora-

tion of these results can and should be pursued. A next

logical step would be to combine aptitude variables to

describe complex interactions, using multiple rather than

simple regression analyses.

Additional methodological issues need also to

be considered. While significant disordinal aptitude x

treatment interactions have been obtained, the points

along the ability continuum at which the effects of in-

structional procedures clearly become significantly dif-

ferent have not yet been established. One possibility

for such analysis lies in the application of the Johnson-

Neyman (1936) method in which the continuum of aptitude

test scores can be divided into regions where each in-

structional method is superior.

Moreover, optimal procedures for dealing with

multiple instructional outcomes have not yet been determined.

Simultaneous statistical consideration of multiple out-

comes of instruction is infrequently represented in the

literature. In view of the differential importance of

multiple criteria, it is not yet apparent what weighting

schemes might be most useful when combination is desirable.

It is anticipated that additional analysis of

the present data will continue in an attempt to provide

information useful in resolving these issues.

APPENDICES

143

APPENDIX A

RATER MANUAL

144

145

RATER MANUAL

I. GENERAL RULES

A. When a teacher is eliciting a student response, he is asking aquestion.

B. Rate each inquiry as given. Don't add punctuation or extra wordseven if this seems necessary to "make sense" out of the inquiry.

If inquiries have blanks or data indicating something has beenomitted, rate anyway as long as the inquiry makes sense.

D. Inquiries such as "What else?" "Can you give another example?""Anything else?" "Keep going," "What is he saying?" "What is hegetting at?" "What about?" "How about?" "Any other?" "Is that whathe's saying?" etc., when following an Analytic Question, are con-sidered an extension of that question and are rated as such.

1. An extension takes precedence over any changes of terms.e.g. "What is meant here by garnering glory?" is a definitionalquestion. "Any other reasons?" is considered an extension ofthe definition despite the change in terms.

2. An extension does not refer to a paraphrase, but to the questionpreceding the paraphrase.

E. In case of a conflict of terms use the first term, e.g. "Whatassumptions are made about the reason why people riot?" would berated as an assumption since the assumption precedes the hypothesishere.

F. If a ratable Analytic Question which does not refer specifically tothe written material is embedded in a section of teacher discoursewhich is obviously referring to the article, rate it as Analytic.

G. Questions to be rated as Analytic should be related to the writtenmaterial rather than simply representing the student's opinion.

146

e.g. "What does the author mean by patriotism?" would be rated asAnalytic, whereas "What is your on definition of patriotism?"would not be.

H. Questioning should be directed to critical analysis of the writtenmaterial. This should be interpreted fairly liberally, with thebenefit of the doubt given in borderline cases. The main distinc-tion is that the teacher should be getting at the analysis of thematerial--what has actually been set forth--rather than simplyasking students for their on opinions, value judgments, etc.

I. Be sure to check to see if the teacher discourse at the 12,2 of a

page may be a continuation from the bottom of the preceding page- -

this can make a big difference as to whether the question is to berated as an Analytic Question or not.

J. Be sure to check all of your totals to see that ALL questions havebeen rated. This is VERY important.

II. ANALYTIC QUESTIONS

A. Hypotheses

A hypothesis states a relation of dependency among variables.It predicts that s change in one variable will be accompanied by or willproduce changes in another variable; thus the relationship may be statedas cause and effect or as accompaniment.

Here are some examples of hypotheses: The characteristics ofa college a student attends will be related to the scores he obtains onthe GRE. The first variable is "college characteristics"; the second"examination scores." Attitude toward school will be related positivelyto the amount of biology retained. Attitude toward school is onevariable (since not all students have the same attitude), amount ofbiology retained is another variable. The hypothesis predicts thatpositive attitudes will be associated with greater retention.

Hypotheses are frequently (but not always) stated in the form:If A, then B: If positive attitude toward school, then greater reten-tion of subject matter.

pP

tryry

147

Includes: Any interrogative sentences or statements that ask what

relationship exists between two or more variables.

Examples

1. What is the author's hypothesis?2. What is the problem that is being investigated?3. What does he think is the cause of B?

4. What does he think A causes?5. What is thought here to be the relationship between A and B?

6. How are A and B connected?7. Is X the hypothesis?8. What analogy is he stating?9. Why does he think A happens?

10. What arguments (explanation) does he give for B?

11. How does he account for B?12. What is he trying to prove?

BOTH SIDES OF THE HYPOTHESIS HAVE TO BE RELATED TO THE ARTICLE.

Does not include: What is theme, or main point.

B. Definitions

The word definition comes from the Latin de which means,

"concerning" and "finis," boundary or limits. A definition is therefore

literally a statement concerning the limits or boundaries of the meaning

of a word. Definitions thus have two main tasks to perform: to convey

the essential meaning which is to be the common ground of understanding,

and to mark out its limits with sufficient precision for the purposes

in view.

Includes: Questions requiring students to distinguish the methods and

standards used for defining a word in terms of the written material

they are analyzing (as distinguished from their own opinion or common

sense notion), similarities, differences, definitions by example.

Examples

1. What is meant here by poverty?

2. A proposed civil rights bill would apply to business engaging

in a "substantial" amount of interstate commerce. How is

substantial defined here?

148

3. What is meant here by patriotism?4. How do we know what a hippie is from reading this?5. How does he define objective?6. Does he define :hippie?7. What distinctions does he make between militaristic and

pacifistic patriotism?8. What does he say is the difference between prejudice and

discrimination?9. Who are the hippies? (who here referring to a collective)

10. How does the author define poverty with respect to incomelevel?

11. What examples does he give of patriotism?

Does not include: "Who" when who is an individual rather than collec-tive. Describe is not included unless the question is something like"describe what he means by IiippiJ77-3T unless words other than describeare added to give it additional meaning.

C. Assumptions

Includes questions requiring students to identify assumptionsbeing made which can only be inferred from an analysis of a series ofstatements.

Includes: Questions asking for the identification of an element re-garded as that which is assumed, inferred, implied, presupposed, takenfor granted, underlying, behind, underneath, etc.

Examples

1. What assumptions are being made in this argument?2. What is he inferring about the race riots?3. Is he implying X?1i. What inferences are being made here?5. What is being taken for granted (presupposed) here?6, What must be assumed to underlie that?7. What is behind (underneath) that argument?

149

D. Distinction of Fact from Opinion or Value Judgment

These questions require students to detect the nature andfunction of a particular statement in a communication.

Includes: Questions requiring students to distinguish factual evidencein an argument from opinions or value judgment. (Value judgments referto the good-bad, approve-disapprove, for-against dimension.)

Examples

1. What facts are being presented here?2. What evidence is given in support of the hypothesis?3. How does the author feel about X?Li. What does he think about X?5. What is his idea on this?6. What is his opinion here?7. Is X a fact/ opinion/ value judgment?8. What does he cite as evidence?9. What does he use to defend/back up this hypothesis?

10. What information was that based on?

Does not include: Just because the author said it does not make it afact.

E. Conclusions

Questions requiring students to distinguish a proposed conclu-sion arrived at from supporting statements.

Includes: conclusions, answer, solution, summary point, resolution,final decision arrived at, suggestion made to deal with, alternativesolutions to a problem, ways proposed to deal with a problem, sugges-tions to deal with a problem, etc.

Examples

1. What conclusion does he come to?2. What answer does he arrive at about what the hippies should do?3. Is X the conclusion? (Is that what he is saying? Does he say

that?)

150

4. When he says X, what logical conclusion is he arriving at?5. How does he answer the original question?6. How does he propose to deal with that problem?7. What alternatives to that solution does he offer?

III. QUALITY DIMENSION

A. Identification of Hypotheses - Questions requiring a statement of arelation of dependency among variables.

1. Questions of high quality require students to supply bothvariables involved in the relationship in addition to therelationship itself.

a. What is the hypothesis being irvestigated?b. What is the author trying to prove?c. What relationship is the major premise of his argument?

2. Questions which do not meet the criteria established for highquality include questions in which the student must supply onlyone (or part) of the variables, and/or the relationship itself;

a. What does A cause?b. Why does B happen?c. What is C associated with?d. What is the relationship between A and B?

and questions requiring only that the student agree, disagree,or select from among given alternatives.

a. Is X the hypothesis?b. Is X or Y or Z the hypothesis?

B. Identification of Definitions - Questions requiring students todistinguish the methods and standards used for defining a word.

1. Questions of high quality require students to supply key ele-ments of the definition and their methods of measurement.

a. A proposed civil rights bill would apply to business engag-ing in a "substantial" amount of interstate commerce. Howis "substantial" being defined here?

b. What does the author mean by patriotism?

151

2. Questions which do not meet the criteria established for highquality include questions in which a key element or method ofmeasurement is given;

a. How does the author use the word poverty with respect toincome level?

b. How does the author define patriotism in its military sense?

c. With respect to the amount of tar and nicotine content, howdoes the author define a "revolutionary filter"?


a. Is X what is meant here by revolutionary?b. With respect to social class, does the author mean income,

education, or occupation?

C. Identification of Assumptions - Questions requiring students tosupply both the assumption and its source (what it concerns or is

in regard to).

1. Questions of high quality require students to supply theassumption and its source.

a. What is implicit in his argument?b. What assumption is the author making here?

2. Questions which do not meet the criteria established for high

quality include, questions in which students are directed to

the source of the assumption;

a. What assumption is he making about the nature of man?

b. What inferences are being drawn about what is of primeimportance to minority groups?

and questions requiring only that the student agree, disagree,

or select from among given alternatives.

a. Is he assuming X or Y about Z?b. Is A the assumption being made?

152

D. Distinction of Factual from Nonfactual Statements - Questionsrequiring students to distinguish the nature and function of aparticular statement in a communication as fact, opinion, or valuejudgment.

1. Questions of high quality require students to supply the fact/opinion/value judgment in its entirety.

a. What evidence is presented in support of the hypothesis?b. What facts are given?c. What value judgments are being made?d. What is the author's opinion of that?

2. Questions which do not meet the criterion established for highquality include questions in which part of the relevant infor-mation is given;

a. What facts are cited about the success of current hearttransplants?

b. What value judgment is being made about the lact of validityin the hippie movement?


a. Does he think X about Y?b. Is X a fact?c. Is his opinion X or Y?

E. Identification of Conclusions - Questions requiring students todistinguish conclusions from supporting statements?

1. Questions of high quality require students to supply a conclu-sion in its entirety.

a. What conclusion did the author arrive at?b. What solution to the problem is being suggested?c. What alternatives has he suggested to that answer?

2. Questions which do not meet the criterion established for highquality include questions in which part of the informationrelevant to the conclusion;

a. What conclusion did the author come to on what responsibledissenters should do?

153

b. What answer does the author arrive at as to which is thebetter kind of patriotism?


a. Is the author suggesting that disease prevention is thealternative to heart transplants?

b. Is his solution to the racial problem integration,strengthening of the black community, or both?

APPENDIX B

SET INDUCTION MATERIALS

155

STANFORD UNIVERSITYSchool of Education

Technical Skills Project Prepared by Mary Lou KoranDirector: F. J. McDonald Training Instructions for:

The Development of Analytic Questioning

PLEASE READ THE FOLLOWING INSTRUCTIONS VERY CAREFULLY BEFORE TEACHINGYOUR FIRST LESSON,.

Educators hold that one of the major objectives of instructionin the schools is the development of skills or strategies which will en-able students to become lifelong autonomous learners. While the contentof any field may undergo rapid change, the ability to analyze written ma-terials within that content area is a skill which is always an objectiveof any field of study. One way of developing such ability in students isto train teachers to ask the kinds of questions that require students toengage in analytic thinking with respect to their reading material. Theobjective of this training session is to help you develop the skill ofasking these kinds of questions.

Analytic Questioning

A communication is composed of a large number of elements. Someof these elements are explicitly stated in the communication and can berecognized and classified relatively easily.

There are other elements in a communication which are not soclearly labeled or identified by the writer. These elements can be in-ferred only from an analysis of a series of statements within the communi-cation. Many of these elements may be of great importance in determiningthe nature of the communication. Until the reader can detect them he mayhave difficulty in comprehending and evaluating the communication.

Analytic questioning breaks a communication down into its crit-ical elements or parts such that the function of these elements within thecommunication, as well as the relative hierarchy of ideas, is made clear.Such analyses are intended to clarify the structure of the communication,to indicate how it is organized, and to identify the arrangement of ideasin it.

156

YOUR GOAL IS TO TEACH A LESSON IN SUCH A WAY AS TO ASK QUESTIONS WHICH

LEAD STUDENTS TO IDENTIFY AND CLASSIFY THE CRITICAL ELEMENTS OF THE WRITTEN

COMMUNICATION YOU HAVE BEEN GIVEN.

The written material you have been given is meant to serve as avehicle for analytic questioning. Using the skill of analytic questioningis the objective of the lesson rather than transmitting content.

APPENDIX C

WRITTEN MODEL

157

158

TEACHER: We are dealing with an essay by Archibald McLeish called "When

We Are Gods." You have all had a chance to read this closely now. An

essay, a piece of writing, in fact, any communication can be thought of as

a number of elements together; and usually in a complex interrelationship.So in order to understand the elements that make up a communication, we

need to do some critical analysis. We need to identify these elements,single them out and consider what they are doing in the essay, how they

are contributing to the meaning. So we are going to try to do that with

five categories that relate to the basic thinking process. In fact, this

could be a piece of scientific writing, it could be from a history text

book. As it is, it is an editorial by a philosopher and poet. (NOTICE

THE CATEGORIES OF ANALYTIC QUESTIONS TO BE USED.) The five categories

that we are going to apply to it are: hypotheses, the difference betweenstatements of fact and statements of value, unstated assumptions that he

may be making, definitions and how he is making it known what he means bycertain words, and conclusions that he reaches in the course of his rea-

soning, The categories again are: hypothesis, fact value statements,

basic assumptions, definitions, and conclusions. (NOTICE HOW HE ASKS FOR

HYPOTHESES.) Let's begin with that first category because we could look

for a very broad hypothesis.

STUDENT: I think the idea of the essay is sort of centered around a

paradox that he finds and he brings out in the first paragraph of the rest

of the essay. And the paradox being:, why in this advanced age of ourswith all of our great technological advances, why is man and most of human-

ity still despondent and why does it still feel a great deal of despair

over its own existence.

TEACHER: Okay. He centers then on that paradox. The question seems to

be why the triumph on one side, the uneasiness of humanity on the other.

And notice that he organizes the essay around those two sides: the tri-

umph on the one side, uneasiness on the other. Can you see within the

rest of the essay any place where he seems to be stating another or per-

haps a related hypothesis?

STUDENT: In the third paragraph the first sentence fits nicely into an

"if then" statement. If, in the terms of scientific and industrial accom-plishments, then our age is one of the great ages of history.

TEACHER: Okay. And we can see this as a hypothesis that he goes on to

develop and give some evidence for. (NOTICE HOW HE ASKS FOR THE DISTINC-

TION BETWEEN FACT AND VALUE JUDGMENT.) If we look at it from industrialscientific standpoints, this is one of the great ages of history. Does

he give any facts in relation to this statement? Does he refer to any-

thing that we could call facts rather than value judgmentdor opinions?

159

STUDENT: In the first paragraph, the fact could be the President's state-ment about that nobody seems to see what is right with America, only whatwas wrong with it. A value judgment or a conclusion can be one of thePresident's. "What was right," he said, "was obvious and admirable." Andthen the rest of the essay would probably consist mainly of value judgmentson the part .1f the author.

TEACHER: Okay. Now this is an editorial and I think that is a very impor-tant observation, Debby, that values seem to be throughout it. That he is

making some value judgments. What are some specific value judgments thathe seems to make in the course of the essay?

STUDENT: Well, he makes several through the whole essay. There is one

very blatant one on the second page; it says, "in most ages it is the artswhich are creative and believe; the men of action who despair." And thenhe goes on to say that the arts usually see the truth. And this is apretty extreme assumption, although it may be true.

TEACHER: Okay. The arts see the truth. If that were a statement of factwe would be able to prove it. If it is a statement of value, it would ex-

press something that men believe in. I think we can see that distinction.

(NOTICE HOW HE ASKS FOR DEFINITIONS.) Now, when he gets into talking inthe second half of the essay, notice he is talking about the uneasinessand he uses the term "despair." Does he anywhere in the essay give uswhat we could consider a definition of despair?; What is he talking about

when he says "despair"? Cris?

STUDENT: Uh, I think one probably could say that he says that discoveriesof contemporary literature are old discoveries long since made and thenhe says, the discovery that men each really die and the discovery thatmoral human life is meaningless. He realizes that man is human and thathe realizes that he does possess this technological power, and that he doesrealize that he can destroy himself.

TEACHER: Okay. He has defined despair then as the feeling based on thesediscoveries--old discoveries long since made. Notice him giving examplesthere for definition too because he defines the old discoveries long sincemade. How does he define old discoveries long since made

STUDENT: Well, he gives specific examples of these discoveries.

TEACHER: Right. (NOTICE HOW HE ASKS FOR ASSUMPTIONS.) And that he defines

right within the context as we read what he means. Would you say that overall in his writing that there is any basic unstated assumption, or premisethat he would pretty much hold in mind but not be stating directly;. Think

about that. What do you think? Ken?

160

STUDENT: Possibly the assumption, uh, that the reader is fairly well readin mythology because he brings out several instances, especially near theend, where he brings up Greek Mythology, such as man stealing fire fromthe gods. And also he brings up an analogy of Hercules in the end.

TEACHER: Good. In fact, notice in both cases that he doesn't refer tothe name of the god, the one myth of Prometheus, and he doesn't refer tothe name of Hercules or Heracles. Although he just refers indirectly sowe can conclude that he is assuming some familiarity with myths. Can yousee other basic unstated assumptions within the essay?

STUDENT: Well, kind of, in the next to the last par:4?raph he says, "weknow what we are" and this seems to imply that maybe man has a tendencytowards self destruction, that no matter what happens we'll end up losingourselves.

TEACHER: Good. We know what we are. That almost has a frightening soundto it--we know what we are. (NOTICE HOW HE ASKS FOR CONCLUSIONS.) Whatthen would you say is the conclusion toward which the whole paper movesand would be an answer to the overall problem that he is posing What

is the conclusion?

STUDENT: Well, in the last paragraph i ..is last sentence which would bethe final statement that he makes, he says, like the old Greek here, whowe've already defined as Hercules, who learned when all his labors hadbeen accomplished--that's when all his achievements had been made--thatwould be paralleling Hercules to now, to us, when all our achievementshave been made. That it was he himself who'd killed his sons and that wewould kill our sons, our future with all our achievements--maybe the atombomb or other nuclear weapons.

TEACHER: Good. And that bringing in the myth at the end, you see, espe-cially to someone familiar with the myth, would raise a kind of frighten-ing idea: The old Greek completing his labors;iHercules completing hislabors just like science completes its labors acid then going ins:*,ne andkilling the future, killing his sons. Overall then, we see McLeish rea-soning through, giving in many cases poetic examples. And let's see ifwe can summarize very quickly and see for instance how all that we've said

in a way relates to the hypotheses. What was the major hypothesis that he

dealt with?

STUDENT: Why is there this paradox of great achievement and at the same

time despair.

TEACHER: Right. Why the. . .why the achievement on one side and yet the

almost total despair. And notice how each part of the essay relates to

161

trying to answer that question. Either to justify his statement thatthere has been great triumph in technological scientific areas, or to showthe despair--and then toward the end to answer "Why." And basically, thatanswer, why the despair, is what?

STUDENT: Uhl man is afraid of self-destruction. .

TEACHER: Right.

STUDENT: . . .afraid of his power.

TEACHER: Man is afraid of self-destruction, afraid of his power. .

STUDENT: We know what we are. .

TEACHER: We know what we are. A good sentence to close with--we know whatwe are.

APPENDIX D

WRITTEN MEASURES

163

Below is a list of kinds of questions possible for classroom use. Place

a "T" or "F" in the space provided at the left of each item number

according to whether or not it represents a major kind or category of

Analytic Questioning.

1. Identification of observations

2. Identification of experimental procedures

3. Identification of hypotheses

4. Identification of inferences

5. Identification of theories

6. Identification of conclusions

7. Identification of unstated assumptions

8. Identification of predictions

9. Identification of factual vs. nonfactual statements

10. Identification of manipulation or control of variables

11. Identification of generalizations

12. Identification of semantic definitions

13. Identification of intent of author

14. Identification of the classification of objects or observations1111M110

164

Below is a list of the five major categories of Analytic Questioningfollowed by a list of questions. Identify which of the categories eachquestion fits into by writing the number of the category in the space

provided at the left of the sentence. If a question does not correspond

to any of the given categories, leave the space blank.

1. Identification of hypotheses

2. Identification of factual vs. nonfactual statements

3. Identification of assumptions4. Identification of semantic definition

5. Identification of conclusions

What does the author believe has caused an increase in the use of

narcotics?

Into which categories would Hemingway's writing fall?

What is the writer's opinion on that question?

A proposal for a civil rights bill would apply to businesses en-

gaging in a substantial amount of interstate commerce. What is

meant here by substantial?

Which statement is the logical outcome of his argument?

How could you predict the form of the written material if you knew

its function?

Could you state something that you believe is true for all of these

cases?

What statement constitutes the major premise of the argument?

What would happen if I increased the temperature of the water?

What does the author think is the relationship between poverty and

crime?

What is the answer to that question in your own opinion?

What kind of an experiment could you set up to test the theory

that form determines function?

What inference could you make from these observations?

How does the writer feel about student demonstrations?

What underlying belief is implicit in this argument?

What information would we need to have about that?

What can be inferred in the data from which Galileo extrapolated

the case for free fall?

What examples are given here of "poverty and deprivation"?

165

APPENDIX

RATING FORMS

166

RATING FORMS

TOTAL ANALYTIC QUESTIONS

Hypotheses

Definitions

Assumptions

Fact - Nonfact

Conclusions

TOTAL CATEGORIES

TOTAL HIGH LEVEL

TOTAL NONANALYTIC QUESTIONS

NAME:

DATE:

SCRIPT:

167

NAME:

DATE:

SCRIPT:

ANALYTIC QUESTIONSTotal

HYPOTHESES:i

High Quality

Low Quality

DEFINITIONS:

High Quality

Low Quality

ASSUMPTIONS:

High Quality

Low Quality

FACT - NONFACT:

High Quality

Low Quality

CONCLUSIONS:

High Quality

Low

NONANALYTIC QUESTIONSTotal

APPENDIX F

INTERCORRELATIONS AMONG DEPENDENT VARIABLES

169

INTERCORRELATIONSAMWDEPENDENT VARIABLESFILM-MEDIATED MODELING GROUP

2 3 4 5 6 7 8 9 10 11 12 13 14

1

2

3

4

5

6

7

8

9

10

11

12

13

14


Categories; T1.

High Quality, T1

Nonanalytic, T1


Categories, T2

High Quality, T2

Nonanalytic, T2

Analytic Questions

Categories, T3

High Quality, T3

Nonanalytic, T3

True-False Test

Matching Test

80 83

72

17

15

15

33

11

26

13

20

11

06

-20

53

33

17

29

11

92

54

00

11

-10

43

14

-05

07

13

01

16

08

52

27

44

-04

07

01

09

-17

50

49

50

-08

68

17

04

27

11

50

2

48

-18

91

66

-05

03

-11

54

02

-26

-03

64

-16

-16

-23

07

-07

00

08

22

-08

12

-07

24

07

26

07

16

19

07

06

26

30

20

11

25

12

25

-05

33

Decimals omitted

INTERCORRELATIONS AMONG DEPENDENT VARIABLESWRITTEN MODELING GROUP

2 3 4 5 6 7 8 9 10 11 12 13 14

1

2

3

4

5

6

7

8

9

10

11

12

13

14


Categories, T1

High Quality, T1

Nonanalytic, T1


Categories, T2

High Quality, T2

Nonanalytic, T2


Categories, T3

High Quality, T3

Nonanalytic, T3

True-False Test

Matching Test

78 87

70

12

21

17

48

52

42

29

32

34

29

23

77

54

51

49

20

94

73

-20

-18

-13

00

40

-42

-42

50

41

37

35

69

56

60

-30

38

32

19

15

52

58

44

-23

74

47

37

39

29

67

58

62

-26

91

73

-09

-02

-03

25

24

-33

-29

58

-19

-27

23

14

23

00

18

48

47

46

-22

46

48

46

-07

-37

-17

-40

-11

-04

10

-06

-04

-06

17

-06

-04

49

Decimals omitted

INTERCORRELATIONS AMONG DEPENDENT VARIABLESCONTROL GROUP

2 3 4 5 6 7 8 9 7_0 11 12 13 14

1

2

3

4

5

6

7

8

9

10

11

12

13

14


Categories: T1

High Quality, T1

Nonanalytic, T1


Categories, T2

High Quality, T2

Nonanalytic, T2


Categories, T3

High Quality, T3

Nonanalytic, T3

True-False Test

Matching Test

72 79

61

12

17

18

46

39

27

-17

42

40

31

-12

79

43

35

37

-09

73

66

12

15

02

58

-01

00

-07

70

47

68

12

47

47

46

15

55

39

44

02

53

60

47

07

79

59

42

68

17

28

32

38

17

91

68

37

34

32

49

18

35

25

58

45

43

38

08

-04

-09

16

00

-12

00

18

16

-10

-19

-05

00

-05

-11

-29

03

-18

-12

-03

-03

-01

00

-07

.24

Decimals omitted

APPENDIX G

ANALYSES FOR HOMOGENEITY ASSUMPTION

173

SUMMARY OF REGRESSION ANALYSES BETWEEN TEACHING SESSIONS

Variables

Treatment Group

Film-Mediated

Modeling

ab

Written

Modeling

ab

Control

a

Analytic Questions, T1 -T2

11.10

.64

4.89

.90

1.84

.40

1.05

NS


6.59

.52

2.88

.88

.99

.59

2.14

NS


12.70

.25

5.79

1.18

.17

.74

3.15

.05

Categories of Analytic Questions, T1 -T2

3.55

.12

2.07

.47

.90

.39

1.13

NS

Categories of Analytic Questions, T2 -T3

2.06

.48

1.32

.57

.49

.53

.16

NS

Categories of Analytic Questions, T 1-T 3

3.83

.01

2.22

.44

.65

.39

2.02

NS

High Quality Analytic Questions, T1 -T2

8.49

.74

3.11

1.00

1.19

.33

1.46

NS


5.09

.51

2.98

.72

.70

.51

.64

NS


8.73

.76

4.83

.91

.19

.68

.14

NS

Nonanalytic Questions, T1 -T2

6.04

.43

19.42

-.01

11.0

.57

3.64

.05

Nonanalytic Questions, T2- T3

6.47

.56

8.36

.46

7.74

..72

1.12

NS

Nonanalytic Questions, T1 -T3

4.50

.47

11.07

.25

10.07

..62

1.60

NS

Note:

a and b take the form:

y = a + bx

-4

VARIANCE-COVARIANCE MATRICES

Control Written Modeling Film-Mediated Modeling

ANALYTIC QUESTIONS

T1 T2 T3 T1 T2 T3 T1

T2 T3

T1 9.40 3.35 7.15 9.02 8.17 10.73 11.37 7.25 2.79

T2 3.35 5.65 3.68 8.17 31.28 22.63 7.25 41.06 21.42

T3 7.15 3.68 10.85 10.73 27.63 50.61 2.79 21.42 40.08

CATEGORIES OF ANALYTIC QUESTIONS

T1 T2 T3 T1 T2 T3 T1 T2

T3

T, 1.08 .44 .40 1.03 .48 .45 1.27 .16 .01

T2 .44 1.08 .61 .48 1.93 1.11 .16 1.42 .68

T3 .40 .61 .94 .45 1.11 1.89 .01 .68 1.36

HIGH QUALITY ANALYTIC QUESTIONS

T1 T2 T3 T1 T2 T3 T1 T2 T3

T1 4.96 1.33 3.46 3.60 3.60 3.31 4.11 3.06 3.11

T2 1.33 2.50 1.38 3.60 14.78 10.68 3.06 27.05 13.83

T3 3.46 1.38 5.20 3.31 10.68 19.60 3.11 13.83 30.54

NONANALYTIC QUESTIONS

T1 T3

T T2 T3

T11

T2 1

T T2 T3

Tl 114.89 67.17 68.45 87.16 -1.06 22.04 88.66 38.04 41.94

T2 67.17 116.39 81.78 -1.06 132.05 61.27 38.04 87.16 49.45

T3 68.45 81.78 165.37 22.04 61.27 83.84 41.94 49.45 66.70

APPENDIX H

ANALYSES OF VARIANCE

CLASSROOM PERFORMANCE MEASURES

176


Source Variation df MS

177

Teaching Session 1

Between Groups 2 .16


.02 NS

Teaching Session 2

Between Groups

Within Groups

2 1030.26 39.13 .01

118 26.32

Teaching Session 3

Between Groups 2 1191.01 35.25 .01


178

ANALYSIS OF VARIANCE OF CATEGORIES

OF ANALYTIC QUESTIONS


Teaching Session 1

Between Groups 2 1.34 1.13 NS


Teaching Session 2

Between Groups 2 46.91 31.47 .01


Teaching Session 3

Between Groups 2

Within Groups 118

66.05 47.13 .01

1.40

179

ANALYSIS OF VARIANCE ON HIGH QUALITY

ANALYTIC QUESTIONS

Source of Variation

Between Groups

Within Groups

Between Groups

Within Groups

Between Groups

Within Groups

df MS

Teaching Session 1

2 .90

118 4.25

.22 NS

Teaching Session 2

2 652.46

118 15.01

45.46 .01

Teaching Session 3

2 734.12

118 18.49

39.69 .01

ANALYSIS OF VARIANCE OF NONALYTIC QUESTIONS

180


Teaching Session 1

Between Groups 2 39.73


.41 NS

Teaching Session 2

Between Groups 2 92111 8.35 .01


Teaching Session 3

Between Groups 2 1262.89 12,17


.01

181


Source df MS F

Film-Mediated Modeling Group

TeachingPersonsResidual

Sessions 2 1335.5139 51.8278 20.35

65.62

Written Modeling Group


Sessions 2 442.0039 61.3378 14.79

29.86

Control Group


Sessions 2 6.3039 18.5078 3.93

1.60

p

.01

.01

NS

182

ANALYSIS OF VARIANCE OF CATEGORIES OF ANALYTIC QUESTIONS

Source df MS



Sessions 2 75.6539 1.9278 1.06

70.70 .01



Sessions 2 18.9839 2.9978 .93

20.40 .01

Contol Group


Sessions 2 1.8339 2.1378 .58

3.15 NS

183

ANALYSIS OF VARIANCE OF HIGH

QUALITY ANALYTIC QUESTIONS

Source df MS



Sessions 2 906.2739 33.9178 13.89

65.19 .01



Sessions 2 225.5239 24.4078 6.79

33.16 .01

Control Group


Sessions 2 2.5039 8.7278 2.11

1.18 NS

184

ANALYSIS OF VARIANCE OF NONANALYTIC QUESTIONS

Source df MS



Sessions 2

3978

772.12167.3037.86

20.10 .01



Sessions 23978

550.22155.8573.60

7.47 .01

Control Group


Sessions 2

3978

5.602749557.73

.10 NS

APPENDIX I

TESTS ON MEANS USING NEWMAN-KEULS PROCEDURE

CLASSROOM PERFORMANCE MEASURES

185

ANALYTIC QUESTIONS - TEACHING SESSION 2

186

TreatmentGroups

111.1111. Control 2. Written 3. Film

Mediated

Ordered Means 3.32 8.16 13.40

Differences be-tween pairs

1.

2.

4.84** 10.08**

5.24**

SA= .77 r

2r3

q .95 2.80 3.36

q .99 3.70 4.20

SA-cl 952.15 2.58

S-107.q .99 2.85 3.23

SOMIIIMM=141MIIINIMEMMEMOMMAIMMIM=Y1

*p < .05

**p < .01

TreatmentGroups


.IIaMIMmMMIMIMII=MMtMms1. Control

Ordered Means 2.93


1.

187

2. Written 3. FilmMediated

10.09 13.58

7.16** 10.65**

3.49**

SA= .77 r

2r3

q .95 2.80 3.36

q .99 3.70 4.20

SAq .95 2.15 2.58

SA q .99 2.85 3.23

*p < .05

**p < .01

188

CATEGORIES OF

ANALYTIC QUESTIONS - TELCHING SESSION 2

IMIIMMIMIIMMI=1.1 .1111.11111=01=PMWMIIIMMIM

Treatment 1. Control 2. Written 3. FilmGroups Mediated

Ordered Means 1.59 2.89


1.

2.

1.30**

3.73

2.14**

.84**

SA

= .18 r2

r3

q .95 2.80 3.36

q .99 3.70 4.20

Szi .95 .40 .60A

SA q .99 .66 .75

.111WMOOMOOM401...........1.. .=.0*p < .05

**p < .01

189

CATEGORIES OF


TreatmentGroups

.11111.111M.IND ONO

1. Control 2. Written 3. FilmMediated

Ordered Means 1.33 2.98


1.

2.

1.65**

3.84

2.51**

.86**

SA= .18 r

2 3

q .95 2.80 3.36

q .99 3.70 4.20

Sq .95 .40 .60A

SA

.99 .66 .75.*p < .05

**p < .01

IIIINNIIINII=

190

HIGH QUALITY


TreatmentGroups

1. Control

im1M.M.I.W.N111m...W11.14.11MiMIIMASOM.1.11111IIIMI.

2. Written 3. FilmMediated



1.

2.

3.26** 7.98**

4.72**

SI= .57 2r3

q .95 2.80 3.36

q .99 3.70 4.20

SA 4 .95 1.60 1.92

Sq .99 2.11 2.39A

*p < .05

(.1) < .01

191

HIGH QUALITY


Treatment 1. Control 2. Written 3. FilmGroups Mediated

Ord d Means 1.68 6.73 10.15


1. 5.05** 8.47**

2. 3.42**

S = .57A

r2 r3

q .95 2.80 3.36

q .99 3.70 4.20

.95 1.60 1.92A

.99 2.11 2.39A

*p < .05

**p < .01

{.

NONANALYTIC QUESTIONS - TEACHING SESSION

192

2

TreatmentGroups

1. FilmMediated

2. Written 3. Control

Ordered Means 16.02 19.13 25.40

1.Differences be-tween pairs

2.

3.11 9.38**

6.27**

SA= 1.61 r

2r3

q .95 2.80 3.36

q .99 3.70 4.20

SAq .95 4.51 5.40

SAq .99 5.96 6.76

*p < .05

**p < .01

"T.

193

NONANALYTIC QUESTIONS - TEACHING SESSION 3

TreatmentGroups

1. Film 2.

MediatedWritten 3. Control

Ordered Means 15.51 16.24 25.96


1.

2.

1.73 10.45**

8.72**

SA = 1.61 r2 r3

q .95 2.80 3.36

q .99 3.70 4.20

SA q .95 4.51 5.40

SA q .99 5.96 6.76

*p < .05

**p < .01

APPENDIX J

TESTS ON MEANS USING NEWMAN-KEULS

PROCEDURE WRITTEN MEASURES

194

MATCHING TEST

195

TreatmentGroups




1.

2.

2.74** 3.68**

.94*

11111101110 .33 r2 r

3

q .95 2.80 3.36

q .99 3.70 4.20

Sq .95 .92 1.10A

S-41 .99 1.22 1.40

*p < .05

**p < .01

TreatmentGroups

196

TRUE-FALSE TEST




1. 2.77** 3.93**

2. 1.16

ellne011111S.44

A 2 3

q .95 2.80 3.36

q .99 3.70 4.20

Sxq .95 1.23 1.48

STiq .99 1.60 1.85

*p < .05

**p < .01

APPENDIX K

INTERCORRELATIONS OF ABILITY MEASURES

197


FILM-MEDIATED MODELING GROUP

12

34

56

78

910

11

12

13

14

15

16

17

18

19

20

21

22

23

1.


59

90

43

32

39

40

43

45

32

08

18

03

13

10

31

-13

25

10

19

07

15

13

2.


87

31

30

33

29

38

37

22

09

16

1P

23

22

24

Ott

19

02

13

07

30

22

3.


42

35

40

39

46

46

31

09

20

11

20

18

31

-05

25

08

18

08

25

20

4.


65

83

41

49

48

24

13

20

32

33

34

32

-05

22

37

34

07

14

12

5.


93

40

48

48

09

18

15

21

34

29

25

00

07

23

17

-02

09

04

6.


46

54

54

20

19

22

29

37

34

32

-01

12

26

22

-01

09

05

7.


72

91

18

02

12

25

25

26

40

09

14

17

21

CO -03

-02

8.


93

36

26

35

37

25

33

36

00

-01

17

10

-15 -13

-15

9.


30

16

27

34

27

32

41

04

05

18

16

-09 -09

-10

10.


68

88

46

29

39

33

10

24

29

28

08

15

14

11.


93

53

40

48

36

00

09

27

22

-03

12

06

12.


56

39

49

39

04

17

30

28

04

14

11

13.


82

94

67

00

21

28

32

23

23

25

14.


95

69

00

30

26

35

33

38

39

15.


72

08

27

28

35

29

32

34

16.

Verbal GRE

00

26

22

32

31

25

29

17.

Film Memory

28

41

35

13

11

13

18.


69

91

59

67

69

19.


89

40 44

46

20.


55

60

63

21.


I68

88

22.


II

94

23.


Decimals omitted.

INTERCORRELATIONS OF ABILITY<I4EASURES:

WRITTEN MODELING GROUP

12

34

56

78

910

11

12

13

14

15

16

17

18

19

20

21

22

23

1.


50

86

18

36

29

26

32

Z1

25

14

20

19

24

22

17

13

02

-04

01

-19

02-14

2.


87

27

48

42

14

20

19

26

15

24

20

14

19

08

32

25

13

25

-06

-01

-09

3.


26

49

41

23

30

29

29

17

25

23

22

23

14

26

16

05

15

-14

00

-13

4.


75

90

33

31

34

26

25

25

35

26

33

19-05

37

21

35

31

21

26

5.


93

41

42

45

21

27

24

37

32

37

27

11

29

15

26

22

12

15

6.


43

41

45

26

29

28

42

31

39

30

02

37

18

33

26

17

21

7.


72

91

10

32

22

11

28

23

31

-03

02

24

67

17

25

21

8.


94

09

23

19

26

34

33

22

01

-04

08

-01

23

26

24

9.


10

29

22

20

34

31

28

02

-01

16

03

21

28

25

10.


73

90

-07

-01

-05

06

18

15

11

19

13

02

05

11.


94-02

07

01

07

04

18

06

-16

16

10

10

12.


-06

03

-02

06

11

15

06

15

15

07

09

13.


77

93

39-15

00

03

-03

25

-02

09

14.


93

42

-01

-03

06

-04

32

13

22

15.


43

-08

00

06

-02

29

08

18

16.

Verbal GRE

-06

01

03

00

25

31

36

17.

Film Memory

17

20

30

-13

-12

-16

18.


32

92

20

12

16

19.


59

22

-03

07

20.


23

07

14

21.


I43

84

22.


II

81

23.


Decimals omitted.


CONTROL GROUP

12

34

56

78

910

11

12

13

14

15

16

17

18

16

20

21

22

23

1.


73

93

31

30

33

27

09

19

18

05

11

-02

08

03

00-15

-05

-20

-13

13

-05

02

2.


92

41

51

49

30

28

32

24

15

18

-13

02

-04

-02

-05

-04

00

-02

-01

-09

-07

3.


39

43

44

31

19

27

22

11

16

-08

05

00

00

-11

-05

-11

-08

06

-07

-02

4.


74

92

40

49

48

12

15

15

11

34

26

37

-02

00

03

00

15

27

26

5.


92

48

48

51

-07

11

06

17

33

28

27

07

-05

10

01

11

22

21

6.


47

52

53

-01

15

06

15

38

30

34

05

00

10

04

14

29

27

7.


80

94-05

-02

-03

-10

11

01

29

42

25

16

25

10

08

11

8.


94

01

11

08

-19

02

-07

20

37

25

16

25

04

03

04

9.


-02

04

01

-15

06

-03

26

41

27

17

26

06

05

07

10.


66

89-04

00-02

08

12

37

01

25

00

02

01

11.


91

01

06

04

14

14

41

23

39

-07

00

-03

12.


-01

06

03

16

14

43

14

35

-01

03

01

13.


75

92

50

13

-01

-01

-01

25

54

50

14.


95

68

25

00

-01

00

32

46

47

15.


64

21

-01

-01

-01

31

53

52

16.

Verbal GRE

23

02

12

07

27

41

41

17.

Film Memory

46

44

52

08

17

16

18.


50

91

-07

14

06

19.


81

02

00

01

20.


-03

10

05

21.


I42

77

22.


II

90

23.


Decimals omitted.

O O

BIBLIOGRAPHY

.

a

201

z.

BIBLIOGRAPHY

Allison, R. B., Jr. Learning parameters and human abili-ties, Princeton, N.J.: Educational TestingService, 1960.

Atkinson, J. W. The psychology of motivation. Princeton:Van Nostrand Co., Inc., 1964.

Bandura, A. Behavior modification through modeling proce-dures. In Krasner and Ullman (Eds.) Researchin behavior modification. New York: Holt, Rine-hart and Winston, 1964.

Bandura, A. Social learning through imitation. Nebraskasymposium on motivation. In. M. K. Jones, ed.,Lincoln: U. of Nebraska Press, 1962, 211-269.

Bandura, A..; and Mischel, W. Modification of self imposeddelay of reward through exposure to live andsymbolic models. J. of pers and soc Psychol,1965, 2:698-705.

Bandura, A., Ross, D., and Ross, S. Imitation of silmmediated aggressive models. J. of ab. and soc.Psychol., 1963, 6, 3-11.

Bandura, A. and Walters, R. Social learning and person-ality development, Chicago: Holt, Rinehart andWinston, 1963.

Blaine, D. D., and Dunham, J. L. Memory abilities in con-cept learning. Paper presented at the meetingof the Southwestern Psychological Association,New Orleans, April, 1968.

Bloom, B. S. (Ed), Taxonomy of educational objectives,New York: Longmans, Green, 1956.

Botzum, W. A. A factorial study of the reasoning andclosure factors. Psychometrika, 1963, 16, 361-386.

Bunderson, C. V. The computer and instructional design.Paper presented at the Conference on ComputerAssisted Instruction, University of Texas, Octo-ber, 1968.

202

203

Carroll, J. B. Study of language. Cambridge, Mass.:Harvard University Press, 1959.

Cronbach, L. J. The two disciplines of scientific psy-chology. American Psychologist, 1957, 12, 671-684.

Cronbach, L. J. Essentials of psychological testing. NewYork: Harper & Row, 1963.

Cronbach, L. J. How can instruction be adapted to indi-vidual differences. In Gagne, R., Learning andindividual differences. Columbus, Ohio: CharlesE. Merril Books Inc., 1967.

Cronbach, L. J. "Individual differences in learningability as a function of instructional variables,"Basic and Applied Research Proposal, Submittedto the U.S. Commission of Education, August 15,1965.

Dixon, W. J. (Ed.), Biomedical computer programs. LosAngeles: School of Medicine, University ofCalifornia, Los Angeles, Revised September,1965.

Duncanson, J. P. Intelligence and the ability to learn.Princeton, N.J.: Educational Testing Service,1964.

Edgerton, H. E. The interaction of trainee aptitude pat-tern with method of instruction. New York:Richardson, Bellows, Henry and Co., 1958.

Fitts, P. M. Perceptual experiments and process theory.In Gagnd, R., Learning and individual differ-ences. Columbus, Ohio; Charles E. Merril BooksInc., 1967.

Fleishman, E. A. The description and prediction of per-ceptual-motor skill learning. In R. Glaser(ed.) Training research and education. NewYork: John Wiley & Sons, Inc., 1965.

204

French, fJ Ekstrom, R., and Price, L. L Manual for kit ofreference tests for cognitive factors. (Revised1963), Princeton, New Jersey: Educational Test-ing Service, 1963.

Gagn41 R. Problem solving in A. W. Melton (Ed.) Cate-zories of human learning, New York: AcademicPress, 1964, p. 293-317.

Gagn4, R. M., and Paradise, N. E Abilities and learningsets in knowledge acquisition. Psychol. Monogr.1961, 75, No. 14 (Whole No. 518).

Glaser, R. Some implications of previous work on learningand individual differences. In Gagne, R., Learn-ing and individual differences. Columbus, Ohio:Charles E. Merril Books, Inc., 1967.

Grimes, J. W. and Allinsmith, W. Compulsivity, anxietyand school achievement. Merril Palmer Quarterly,1961, 7, 247-271.

Humphreys, L. G. Some investigations of the simplex.Psychometrika, 1960, 25, 313-323.

Husband, R. W. Intercorrelations among learning abilities:I, J. genet. Psychol., 1939, 55, 353-364.

Husband, R. W. Intercorrelations among learning abilities:III, The effect of length of tests. J. ofgenet. Psychol., 1941, 58, 427-430.

Johnson, P. 0., and Neyman, J. Tests of certain linearhypotheses and their application to some educa-tional problems. Statis. res Mem, U. of London,1936, 1, 57-93.

Koran, J. J. The relative effects of imitation versusproblem solving on the acquisition of a complexteaching skill, an unpublished dissertation,Stanford University, 1968.

Lloyd Morgan, C. Habit and Instinct. London: Arnold, 1896.

Manning, W. H., and Dubois, P. H. Correlational methodsin research on human learning. Perceptual andMotor Skills, 1962, 15, 287-321, MonographSupplement 3-V15.

205

McBrearty, J. F., Marston, A. F., and Kanfer, F. H. Con-ditioning a verbal operant in a group setting:direct versus vicarious reinforcement. AmericanPsychologist, 1961, 16, 425-434.

McDonald, F. J. and Allen, D. W. Training effects offeedback and modeling procedures on teachingperformance. Stanford University, 1967.

McGuire, W. J. A multiprocess model for paired-associate learning. J. Exp. Psychol., 1961,62:335-347.

Melton, A. W. Individual differences and theoreticalprocess variables: General comments on theconference. In Gagn4, R. (ed.) Learning andIndividual Differences. Columbus, Ohio: CharlesE. Merril Books, Inc., 1967.

Miller, N. E. and Dollard, J. Social learning and imitation.New Haven: Yale University Press, 1941.

Mowrer, O. H. Learning theory and the symbolic processes.New York: Wiley, 1960.

Pemberton, C. L. The closure factors related to othercognitive processes. Psychometrika, 17, 267-288.

Porter, D. Some effects of year long teaching machineinstruction, in E. Galanter (Ed.) AutomaticTeaching. New York: Wiley.

Rapier, J. L. Measured intellifrence and the ability tolearn. Acta Psychologica, 1962, 20, 1-17.

Sanders, N. Classroom questions. New York: Harper andRowe, 1966.

Seibert, W. F., and Snow, R. E. Studies in cine-psychometryI. Final Report, July, 1965, USOE, Grant Number7-12-0280-184.

Seibert, W. F., Reid, C., and Snow, R. Studies in cine-psychometry II. Final Report, December, 1967,USOE, Grant Number 7-24-0880-257.

206

Sheffield, F. D. Theoretical considerations in the learn-ing of complex sequential tasks from demonstra-tion and practice. In A. A. Lumsdaine (Ed.).Student res onse in ro rammed instruction: Asymposium. Wastangton D.C.: National Academyof Sciences--National Research Council, 1961,pp. 13-32.

Simrall, D. Intelligence and the ability to learn. Journalof Psychology, 1947, 28, 27-43.

Snow, R. E., Tiffin, J., Siebert, W. Individual differ-ences and instructional film effects, J. of ed.Psychol., 1965, 56, 315-326.

Stake, R. E. Learning parameters, aptitudes and achieve-ment. Psychometric Monographs. PsychometricSociety, 1961, No. 9.

Tarde, G. The laws of imitation. New York: Holt, 1903.

Thurstone, L. L. A factor analysis study of perception.Psychometr. Monogr., 1944, No. 4.

Wallach, M., and Kogan, N. Risk Taking. New York: Holt,Rinehart & Winston, Inc., 1964.

White, B. W. Visual and auditory closure, J. exp. Psychol.,1954, 48, 234-243.

Winer, B. J. Statistical principles in experimental de-sign. New York: McGraw-Hill, 1962.

Witkin, H. A., Dyk, Ruth B., Faterson, Hanna F., Goodenough,D. R., and Kark, S. A. psychological differen-tiation: Studies of development. New York:Wiley, 1962:

Woodrow, H. The relation between abilities and improve-ment with practice. J. of educ. Psychol., 1938,29, 215-238.

Woodrow, H. Interrelation of measures of learning. J.Psychol., 1940, 10, 49-73.

A.

Woodrow, H. Factors ln improvement with practice. J.Psychol., 1939, 7, 55-70.

Woodrow, H. The ability to learn. Psych. Rev., 1946,53, 147-158.

207

kaArk Anep typing service

2707 HEMPHILL PARK AUSTIN, TEXAS 78705 AC 512 472-3210

II een fi YES 0 NO El 3 I 0 IC CI AUTHOR SOURCE CODE ...Mar 69 II een fi 3 rbil 1 E. N.) fi \ii...

Documents

Transcript of II een fi YES 0 NO El 3 I 0 IC CI AUTHOR SOURCE CODE ...Mar 69 II een fi 3 rbil 1 E. N.) fi \ii...